IB Maths AI notes › Glossary

Glossary

Key terms in plain words, with the lesson that teaches each one.

6

68–95–99.7 rule
For a normal distribution, approximately 68%68\%, 95%95\% and 99.7%99.7\% of values lie within 11, 22 and 33 standard deviations of the mean.Taught in 4.9 Normal distribution

A

acceleration
The rate of change of velocity with time, a=dvdt=d2sdt2=vdvdsa = \frac{\mathrm{d}v}{\mathrm{d}t} = \frac{\mathrm{d}^{2}s}{\mathrm{d}t^{2}} = v\frac{\mathrm{d}v}{\mathrm{d}s}; in straight-line motion, a negative acceleration slows an object down only while its velocity is positive, because speed falls only when velocity and acceleration have opposite signs.Taught in 5.13 Kinematics
adjacency matrix
A square matrix with one row and one column for each vertex of a graph, whose entry in row ii, column jj is the number of edges going directly from vertex ii to vertex jj; it is symmetric when the graph is undirected.Taught in 3.15 Adjacency matrices and walks
alternative hypothesis
The statement H1\mathrm{H}_1 that a test finds evidence for when it rejects H0\mathrm{H}_0; its form (≠\neq, or >> or <<) decides whether a t-test is two-tailed or one-tailed.Taught in 4.11 Chi-squared tests and the t-test
ambiguous case
When you know two sides of a triangle and an angle not between them, the sine rule can give both an acute and an obtuse angle, so two, one or no triangles may fit; keep each angle whose sum with the known angle is less than 180∘180^{\circ}.Taught in 3.8 Unit-circle trigonometry and equations
amortization
Repaying a loan with equal regular payments, each covering that period's interest with the rest reducing the amount owed, so the debt reaches 0 with the last payment; in IB exams the payments are made at the end of each period.Taught in 1.7 Amortization and annuities
amplitude
For a sine or cosine wave, half the difference between its maximum and minimum values; for asin⁡(bx)+da\sin(bx) + d or acos⁡(bx)+da\cos(bx) + d it is ∣a∣|a|, which is never negative.Taught in 2.5 Modelling with functions
angle between a line and a plane
For a line that meets a plane, the angle between the line and its projection on the plane, the line joining the meeting point to the foot of a perpendicular dropped from a point on the line to the plane; it is never more than 90∘90^{\circ}.Taught in 3.1 3D distance, midpoint and measurement
angle of depression
The angle between the horizontal (not the vertical) and the line of sight down to an object below the observer; over horizontal ground it equals the angle of elevation from the object back to the observer.Taught in 3.3 Applied triangle trigonometry
angle of elevation
The angle between the horizontal and the line of sight up to an object above the observer.Taught in 3.3 Applied triangle trigonometry
annuity
A lump sum invested now that pays out a fixed amount each period for a set number of periods, while the money left in it keeps earning compound interest; in IB exams the payments are made at the end of each period.Taught in 1.7 Amortization and annuities
antiderivative
A function FF whose derivative is ff, so F′(x)=f(x)F'(x) = f(x); on an interval, any two antiderivatives of ff differ only by a constant.Taught in 5.5 Antidifferentiation, definite integrals and area
arc
Part of the circumference of a circle; with radius rr and angle θ\theta degrees at the centre, its length is θ360×2πr\frac{\theta}{360} \times 2\pi r.Taught in 3.4 Arc length and sector area
Argand diagram
A drawing of the complex plane, with a horizontal real axis and a vertical imaginary axis, on which a+bia + b\mathrm{i} (aa, bb real) is the point (a,b)(a, b).Taught in 1.12 Complex numbers in Cartesian form
argument of a complex number
For z=a+bi≠0z = a + b\mathrm{i} \neq 0, the angle arg⁡z\arg z from the positive real axis to the line from 00 to zz on an Argand diagram, measured anticlockwise and usually with −π<arg⁡z≤π-\pi < \arg z \le \pi; arctan⁡ba\arctan\frac{b}{a} gives it only when a>0a > 0, so check the quadrant.Taught in 1.12 Complex numbers in Cartesian form
arithmetic sequence
A sequence in which each term is the previous term plus the same constant dd (positive, negative or zero), so the terms rise, fall or stay level in equal steps.Taught in 1.2 Arithmetic sequences and series
asymptote
A straight line that a graph gets closer and closer to as xx or yy grows without limit, written as an equation such as x=3x = 3 (vertical) or y=2y = 2 (horizontal); a graph may cross a horizontal asymptote at a finite value of xx.Taught in 2.4 Key features of graphs
axis of symmetry
A line that a graph can be folded along so the two halves match; for the parabola y=ax2+bx+cy = ax^{2} + bx + c it is the vertical line x=−b2ax = -\frac{b}{2a} through the vertex.Taught in 2.4 Key features of graphs

B

base vectors
The unit vectors i\boldsymbol{i}, j\boldsymbol{j} and k\boldsymbol{k}, one unit along the positive xx-, yy- and zz-axes, so that any vector can be written v1i+v2j+v3kv_{1}\boldsymbol{i} + v_{2}\boldsymbol{j} + v_{3}\boldsymbol{k}.Taught in 3.10 Vectors in 2D and 3D
bearing
A direction given as the angle measured clockwise from north, written with three figures before any decimal point, such as 045∘045^{\circ} for north-east.Taught in 3.3 Applied triangle trigonometry
binomial distribution
The distribution of the number of successes XX in a fixed number nn of independent trials, each with two outcomes and the same probability pp of success; it is written X∼B(n,p)X \sim \mathrm{B}(n, p) and has mean npnp and variance np(1−p)np(1-p).Taught in 4.8 Binomial distribution
bivariate data
Data that comes in pairs: two measurements, such as height and mass, taken from each member of a sample.Taught in 4.4 Correlation and regression (y on x)
boundary condition
A known pair of values, such as y=10y = 10 when x=1x = 1, substituted into the integrated function (not the derivative) to find the constant of integration.Taught in 5.5 Antidifferentiation, definite integrals and area
box and whisker diagram
A diagram drawn to scale with a box from Q1Q_1 to Q3Q_3, a line at the median, whiskers out to the smallest and largest values that are not outliers, and a cross for each outlier.Taught in 4.2 Data presentation

C

carrying capacity
In a logistic model f(x)=L1+Ce−kxf(x) = \frac{L}{1 + C\mathrm{e}^{-kx}} with L,C,k>0L, C, k > 0, the value LL: the horizontal asymptote y=Ly = L, a long-term limit the quantity approaches but never reaches.Taught in 2.9 Further function models
Cartesian form
A complex number written as z=a+biz = a + b\mathrm{i} with aa and bb real: the real part first, then the imaginary part times i\mathrm{i}.Taught in 1.12 Complex numbers in Cartesian form
causation
When a change in one variable directly produces a change in another; a correlation alone, however strong, does not show causation.Taught in 4.4 Correlation and regression (y on x)
cell (Voronoi diagram)
The region of all points closer to one site than to any other site; its edges lie on perpendicular bisectors of that site and its neighbouring sites.Taught in 3.6 Voronoi diagrams
central limit theorem
The result that the mean Xˉ\bar{X} of a random sample of size nn from any population with mean μ\mu and variance σ2\sigma^{2} is approximately N(μ,σ2n)\mathrm{N}\left(\mu, \frac{\sigma^{2}}{n}\right) when nn is large, where how large depends on the population and IB exams treat n>30n > 30 as enough.Taught in 4.15 Sample mean and the central limit theorem
chain rule
The rule for differentiating a composite function: if y=g(u)y = g(u) and u=f(x)u = f(x), then dydx=dydu×dudx\frac{\mathrm{d}y}{\mathrm{d}x} = \frac{\mathrm{d}y}{\mathrm{d}u} \times \frac{\mathrm{d}u}{\mathrm{d}x}, so you differentiate the outside function and multiply by the derivative of the inside.Taught in 5.9 Differentiation rules and related rates
characteristic polynomial
For a 2×22 \times 2 matrix AA, the quadratic det⁡(A−λI)\det(A - \lambda I) in λ\lambda, whose roots are the eigenvalues of AA.Taught in 1.15 Eigenvalues and eigenvectors
Chinese postman problem
Finding the least-weight closed route along every edge of a connected weighted graph at least once, by repeating the shortest routes for the least-weight pairing of the odd vertices.Taught in 3.16 Spanning trees, postman and salesman problems
circuit
A trail that starts and finishes at the same vertex, so it is closed and repeats no edge.Taught in 3.16 Spanning trees, postman and salesman problems
class interval
A group of values in a grouped frequency table, written as an inequality such as 20≤t<2520 \le t < 25; the classes leave no gaps and don't overlap, so each value belongs to exactly one class.Taught in 4.2 Data presentation
coefficient of determination
R2R^2, the proportion of the variation in yy that a model accounts for; it equals r2r^2 for a linear least squares model, and a high R2R^2 on its own does not show that a model is the right choice.Taught in 4.13 Non-linear regression and R²
common difference
In an arithmetic sequence, the constant d=un+1−und = u_{n+1} - u_{n}, the same for every pair of consecutive terms and negative when the terms fall; for real data that is only roughly arithmetic it can only be estimated.Taught in 1.2 Arithmetic sequences and series
common ratio
The number r=un+1unr = \frac{u_{n+1}}{u_{n}} that every term of a geometric sequence is multiplied by to give the next term; it is the same for every pair of consecutive terms, and it can be negative or a fraction but never 00.Taught in 1.3 Geometric sequences and series
complementary events
An event AA and its complement A′A' (not AA), which holds every outcome not in AA; exactly one of them happens on each trial, so P(A)+P(A′)=1\mathrm{P}(A) + \mathrm{P}(A') = 1.Taught in 4.5 Probability foundations and expected value
complete graph
A simple graph in which every pair of vertices is joined by an edge; the complete graph KnK_n on nn vertices has n(n−1)2\frac{n(n-1)}{2} edges.Taught in 3.14 Graph theory basics
complex conjugate
For z=a+biz = a + b\mathrm{i} with aa and bb real, the conjugate is z∗=a−biz^{*} = a - b\mathrm{i}, the reflection of zz in the real axis of an Argand diagram; the non-real roots of a quadratic with real coefficients are a conjugate pair.Taught in 1.12 Complex numbers in Cartesian form
complex number
A number z=a+biz = a + b\mathrm{i}, where aa and bb are real and i\mathrm{i} is the non-real number with i2=−1\mathrm{i}^{2} = -1; real numbers are the complex numbers with b=0b = 0.Taught in 1.12 Complex numbers in Cartesian form
component of a vector
One of the numbers v1v_{1}, v2v_{2}, v3v_{3} in v=v1i+v2j+v3k\boldsymbol{v} = v_{1}\boldsymbol{i} + v_{2}\boldsymbol{j} + v_{3}\boldsymbol{k}, giving how far the vector goes along the xx-, yy- or zz-axis, written in that order in a column vector.Taught in 3.10 Vectors in 2D and 3D
component of a vector in a direction
For a non-zero vector b\boldsymbol{b}, the component of a\boldsymbol{a} in the direction of b\boldsymbol{b} is the scalar a⋅b∣b∣=∣a∣cos⁡θ\frac{\boldsymbol{a} \cdot \boldsymbol{b}}{\lvert\boldsymbol{b}\rvert} = \lvert\boldsymbol{a}\rvert\cos\theta, where θ\theta is the angle between them (negative when θ>90∘\theta > 90^{\circ}), and the component perpendicular to b\boldsymbol{b}, in the plane of the two vectors, is ∣a×b∣∣b∣=∣a∣sin⁡θ\frac{\lvert\boldsymbol{a} \times \boldsymbol{b}\rvert}{\lvert\boldsymbol{b}\rvert} = \lvert\boldsymbol{a}\rvert\sin\theta.Taught in 3.13 Scalar and vector products
composite function
The function f∘gf \circ g that applies gg first and then ff to the result, so (f∘g)(x)=f(g(x))(f \circ g)(x) = f(g(x)); it exists only where g(x)g(x) lies in the domain of ff, and swapping the order usually gives a different function.Taught in 2.7 Composite and inverse functions
compound interest
Interest that is added to the balance at the end of each period and then earns interest itself; at a nominal annual rate of r%r\% compounded kk times a year, the balance is multiplied by 1+r100k1 + \frac{r}{100k} each period.Taught in 1.4 Compound interest and depreciation
concave-down
Bending downwards on an interval, with the gradient decreasing as xx increases; a graph is concave-down on any interval where f′′(x)<0f''(x) < 0 throughout.Taught in 5.10 Second derivative and concavity
concave-up
Bending upwards on an interval, with the gradient increasing as xx increases; a graph is concave-up on any interval where f′′(x)>0f''(x) > 0 throughout.Taught in 5.10 Second derivative and concavity
conditional probability
The probability of AA given that BB has happened, written P(A∣B)\mathrm{P}(A \mid B) and equal to P(A∩B)P(B)\frac{\mathrm{P}(A \cap B)}{\mathrm{P}(B)}, where P(B)>0\mathrm{P}(B) > 0.Taught in 4.6 Probability diagrams, rules and independence
confidence interval
A range of values for a population mean, calculated from one random sample by a method that gives an interval containing the true mean for a stated percentage of all such samples; a 95% interval does not mean a 95% chance that this particular interval contains the mean.Taught in 4.16 Confidence intervals for the mean
connected graph
A graph in which you can travel along edges from any vertex to any other vertex.Taught in 3.14 Graph theory basics
constant of integration
The + C+\,C at the end of an indefinite integral, there because any constant differentiates to 00; on a domain with no breaks, one known point on the curve fixes its value.Taught in 5.5 Antidifferentiation, definite integrals and area
content validity
The extent to which a test covers every part of what it is meant to measure, in sensible proportions.Taught in 4.12 Data collection, reliability and validity
contingency table
A two-way table of observed frequencies, with one categorical variable in the rows and another in the columns, used in the χ² test for independence.Taught in 4.11 Chi-squared tests and the t-test
continuous data
Measured data that can take any value in a range, such as a time or a mass, even when it is recorded rounded.Taught in 4.1 Sampling, data quality and outliers
continuous function
Informally, a function whose graph has no breaks or jumps on an interval, so it can be drawn there without lifting your pen; a piecewise function is continuous at a joining point when the rules on both sides give the same value there.Taught in 2.9 Further function models
continuous random variable
A random variable that can take any value in an interval, such as a time or a mass; its probabilities are areas under a curve, so P(X=a)=0\mathrm{P}(X = a) = 0 and P(X<a)=P(X≤a)\mathrm{P}(X < a) = \mathrm{P}(X \le a).Taught in 4.9 Normal distribution
convenience sampling
Choosing the members of the population that are easiest to reach, such as the first people you meet; it is quick but likely to be biased.Taught in 4.1 Sampling, data quality and outliers
correlation
How closely the points of a scatter diagram follow a straight line, described by direction (positive, negative or zero) and strength (strong, weak or none).Taught in 4.4 Correlation and regression (y on x)
cosine rule
In any triangle ABCABC, c2=a2+b2−2abcos⁡Cc^{2} = a^{2} + b^{2} - 2ab\cos C, where CC is the angle between sides aa and bb and cc is the side opposite it.Taught in 3.2 Triangle trigonometry
coupled differential equations
A pair of differential equations, one giving dxdt\frac{\mathrm{d}x}{\mathrm{d}t} and one giving dydt\frac{\mathrm{d}y}{\mathrm{d}t}, in which each rate can depend on both xx and yy, so the two must be solved together.Taught in 5.16 Euler's method
criterion-related validity
The extent to which a test's results agree with an established measure of the same thing (the criterion); a strong correlation in the expected direction between the two suggests the test is valid.Taught in 4.12 Data collection, reliability and validity
critical region
The values of a test statistic that lead to rejecting H0\mathrm{H}_0; for a binomial or Poisson test it is the tail with the largest probability under H0\mathrm{H}_0 that is still less than the significance level.Taught in 4.18 Further hypothesis tests and errors
critical value
The boundary value of a test statistic for a given significance level and sample size (number of pairs or degrees of freedom): correlation is significant when ∣r∣\lvert r\rvert is greater than the critical value of rr, and in a χ2\chi^{2} test H0\mathrm{H}_0 is rejected when χ2\chi^{2} is greater than its critical value.Taught in 4.4 Correlation and regression (y on x)
cumulative frequency
The number of data values less than or equal to a given value, a running total of the frequencies; on a cumulative frequency graph it is plotted at each class's upper boundary.Taught in 4.2 Data presentation
cycle
A circuit that repeats no vertex, except that the start vertex is also the end vertex.Taught in 3.16 Spanning trees, postman and salesman problems

D

decreasing function
A function is decreasing on an interval if its output gets smaller as xx gets bigger there, so a<ba < b gives f(a)>f(b)f(a) > f(b); this holds whenever f′(x)<0f'(x) < 0 for every xx in the interval.Taught in 5.2 Increasing and decreasing functions
definite integral
The number ∫abf(x) dx\int_{a}^{b} f(x)\,\mathrm{d}x, found from ff between the limits x=ax = a and x=bx = b; with a<ba < b, it equals the area between the curve and the xx-axis only when f(x)≥0f(x) \ge 0 throughout, because parts below the axis count as negative.Taught in 5.5 Antidifferentiation, definite integrals and area
degree of a vertex
The number of edge ends at a vertex, with a loop counting twice; in any graph the degrees add up to twice the number of edges.Taught in 3.14 Graph theory basics
degrees of freedom
The number of values free to vary once the totals are fixed: (r−1)(c−1)(r - 1)(c - 1) for a contingency table with rr rows and cc columns, and n−1n - 1 for an SL goodness of fit test with nn categories (one fewer for each parameter estimated from the data).Taught in 4.11 Chi-squared tests and the t-test
deleted vertex algorithm
Deleting one vertex, finding a minimum spanning tree of the rest and adding the two least-weight edges from the deleted vertex; the total is a lower bound for the travelling salesman problem, and the largest such bound is the best.Taught in 3.16 Spanning trees, postman and salesman problems
depreciation
The loss of an item's value over time; at a fixed rate of r%r\% a year the value is multiplied by 1−r1001 - \frac{r}{100} each year, so in this model it falls by a smaller amount each year and never reaches 00.Taught in 1.4 Compound interest and depreciation
derivative
The function f′(x)f'(x), also written dydx\frac{\mathrm{d}y}{\mathrm{d}x} when y=f(x)y = f(x), whose value at each xx is the gradient of the curve y=f(x)y = f(x) there and the rate of change of yy with respect to xx.Taught in 5.1 Limits and the derivative as a rate of change
determinant
A number, written det⁡A\det A or ∣A∣|A|, worked out from a square matrix AA: for a 2×22 \times 2 matrix with rows a,ba, b and c,dc, d it is ad−bcad - bc, and AA has an inverse only when det⁡A≠0\det A \neq 0.Taught in 1.14 Matrices, determinants and inverses
diagonalization
Writing a square matrix as M=PDP−1M = PDP^{-1}, with eigenvectors of MM as the columns of PP and the matching eigenvalues, in the same order, on the diagonal of DD (zeros elsewhere), so that Mn=PDnP−1M^{n} = PD^{n}P^{-1}; in IB AI it is done only for 2×22 \times 2 matrices with two different real eigenvalues.Taught in 1.15 Eigenvalues and eigenvectors
differential equation
An equation that contains a derivative, such as dydx=ky\frac{\mathrm{d}y}{\mathrm{d}x} = ky; solving it means finding the function itself.Taught in 5.14 Setting up and separating differential equations
direct variation
A relationship y=axny = ax^{n} with a≠0a \ne 0 and nn a positive integer, read "yy is proportional to xnx^{n}"; then y=0y = 0 when x=0x = 0.Taught in 2.5 Modelling with functions
directed graph
A graph in which each edge has a direction, shown by an arrow, and is travelled only that way; a vertex's in-degree counts edges coming in and its out-degree counts edges going out.Taught in 3.14 Graph theory basics
direction vector
A non-zero vector parallel to a line; any non-zero multiple of it is also a direction vector of that line.Taught in 3.11 Vector equations of lines
discrete data
Data that can take only separate values, usually counts, such as the number of goals in a match.Taught in 4.1 Sampling, data quality and outliers
discrete random variable
A random variable whose possible values can be listed one by one, such as 0,1,2,30, 1, 2, 3. Counts and scores are discrete.Taught in 4.7 Discrete random variables and expected value
discriminant
For ax2+bx+cax^{2} + bx + c with a≠0a \neq 0, the number Δ=b2−4ac\Delta = b^{2} - 4ac; when aa, bb and cc are real, ax2+bx+c=0ax^{2} + bx + c = 0 has two distinct real roots if Δ>0\Delta > 0, one repeated real root if Δ=0\Delta = 0, and no real roots but two complex conjugate roots if Δ<0\Delta < 0.Taught in 1.12 Complex numbers in Cartesian form
displacement
A change in position, given as a vector: from A to B it is AB→=b−a\overrightarrow{AB} = \boldsymbol{b} - \boldsymbol{a}, which depends only on the start and end points and not on the route, so it is not the distance travelled; in one dimension it is a signed distance, such as a position measured from the origin.Taught in 3.12 Kinematics with vectors
distance travelled
The total length of path an object covers between two times, counting motion in either direction, ∫t1t2∣v(t)∣ dt\int_{t_{1}}^{t_{2}} \lvert v(t)\rvert\,\mathrm{d}t; it equals the size of the displacement only if the object never changes direction in that time.Taught in 5.13 Kinematics
distribution of the sample mean
How the mean Xˉ\bar{X} of a random sample of size nn varies from sample to sample: it has mean μ\mu and variance σ2n\frac{\sigma^{2}}{n} (the variance, not the standard deviation, is divided by nn), and it is exactly N(μ,σ2n)\mathrm{N}\left(\mu, \frac{\sigma^{2}}{n}\right) when the population is N(μ,σ2)\mathrm{N}(\mu, \sigma^{2}).Taught in 4.15 Sample mean and the central limit theorem
domain
The set of input values a function is defined for; in a model, the context can limit it further, for example to times t≥0t \ge 0.Taught in 2.2 Function concepts, domain and range

E

e (Euler's number)
The irrational constant e=2.71828…\mathrm{e} = 2.71828\ldots, used as the base of exponential growth and decay models and of the natural logarithm.Taught in 1.5 Exponents and introductory logarithms
edge (graph theory)
A line in a graph joining two vertices, or joining a vertex to itself, in which case it is called a loop.Taught in 3.14 Graph theory basics
edge (Voronoi diagram)
A boundary between two neighbouring cells; it lies on the perpendicular bisector of their two sites, so every point on it is the same distance from both.Taught in 3.6 Voronoi diagrams
eigenvalue
A number λ\lambda such that Av=λvA\boldsymbol{v} = \lambda\boldsymbol{v} for some non-zero vector v\boldsymbol{v}, where AA is a square matrix; the eigenvalues are the solutions of det⁡(A−λI)=0\det(A - \lambda I) = 0.Taught in 1.15 Eigenvalues and eigenvectors
eigenvector
A non-zero vector v\boldsymbol{v} with Av=λvA\boldsymbol{v} = \lambda\boldsymbol{v} for a square matrix AA and some number λ\lambda, so the matrix only stretches, shrinks or flips it; any non-zero multiple of it is also an eigenvector for the same λ\lambda, and the zero vector never counts.Taught in 1.15 Eigenvalues and eigenvectors
equilibrium point
A point where dxdt=0\frac{\mathrm{d}x}{\mathrm{d}t} = 0 and dydt=0\frac{\mathrm{d}y}{\mathrm{d}t} = 0 at the same time, so a solution that starts there stays there; for dxdt=ax+by\frac{\mathrm{d}x}{\mathrm{d}t} = ax + by, dydt=cx+dy\frac{\mathrm{d}y}{\mathrm{d}t} = cx + dy with non-zero eigenvalues it is the origin only.Taught in 5.17 Phase portraits for coupled equations
Euler's method
A numerical method that estimates the solution of dydx=f(x,y)\frac{\mathrm{d}y}{\mathrm{d}x} = f(x, y) by repeated steps of constant length hh: yn+1=yn+h×f(xn,yn)y_{n+1} = y_{n} + h \times f(x_{n}, y_{n}) and xn+1=xn+hx_{n+1} = x_{n} + h; the result is an approximation, usually closer for a smaller hh.Taught in 5.16 Euler's method
Eulerian circuit
A circuit that uses every edge of a graph exactly once; a connected graph has one exactly when every vertex has even degree.Taught in 3.16 Spanning trees, postman and salesman problems
Eulerian trail
A trail that uses every edge of a graph exactly once; a connected graph has one exactly when it has 0 or 2 odd vertices, and with 2 odd vertices the trail starts at one and finishes at the other.Taught in 3.16 Spanning trees, postman and salesman problems
event
A set of outcomes from the sample space, such as "an odd number"; the event happens when the outcome of the trial is one of them.Taught in 4.5 Probability foundations and expected value
expected frequency
The count predicted for a category if H0\mathrm{H}_0 is true, compared with the observed (recorded) count; in a contingency table it is row total × column total ÷ grand total, and a χ² test needs each one greater than 55.Taught in 4.11 Chi-squared tests and the t-test
expected number of occurrences
The long-run average number of times an event AA happens in nn trials, n×P(A)n \times \mathrm{P}(A); it need not be a whole number.Taught in 4.5 Probability foundations and expected value
expected value
The long-run mean of a discrete random variable XX over many repeats of the trial: E(X)=∑x P(X=x)\mathrm{E}(X) = \sum x\,\mathrm{P}(X = x). It need not be a value XX can take.Taught in 4.7 Discrete random variables and expected value
exponential decay
A decrease in which equal steps in xx multiply the gap f(x)−cf(x) - c by the same factor between 00 and 11, as in kax+cka^{x} + c with k>0k > 0 and 0<a<10 < a < 1, ka−x+cka^{-x} + c with k>0k > 0 and a>1a > 1, or kerx+ck\mathrm{e}^{rx} + c with k>0k > 0 and r<0r < 0; the values approach y=cy = c but never reach it.Taught in 2.5 Modelling with functions
exponential form
A non-zero complex number written z=reiθz = r\mathrm{e}^{\mathrm{i}\theta}, also called Euler form, with the same modulus r=∣z∣>0r = \lvert z\rvert > 0 and argument θ\theta as in polar form, and θ\theta in radians.Taught in 1.13 Complex numbers in polar and exponential form
exponential function
A function with the variable in the exponent, such as f(x)=kax+cf(x) = ka^{x} + c with a>0a > 0, a≠1a \ne 1, k≠0k \ne 0, or f(x)=kerx+cf(x) = k\mathrm{e}^{rx} + c with k≠0k \ne 0, r≠0r \ne 0; its graph has the horizontal asymptote y=cy = c.Taught in 2.5 Modelling with functions
exponential growth
An increase in which equal steps in xx multiply the gap f(x)−cf(x) - c by the same factor greater than 11, as in kax+cka^{x} + c with k>0k > 0 and a>1a > 1, or kerx+ck\mathrm{e}^{rx} + c with k>0k > 0 and r>0r > 0.Taught in 2.5 Modelling with functions
extrapolation
Using a model to predict a value outside the range of data it was built from; less reliable, because the trend may not continue there.Taught in 2.6 Fitting and testing models

F

fair game
A game in which a player's expected gain is 00: E(X)=0\mathrm{E}(X) = 0, where XX is the player's gain (prize minus cost) in one game.Taught in 4.7 Discrete random variables and expected value
fractal
A shape made by applying the same transformations again and again for ever, so it shows detail at every scale; its length or area at each stage often forms a geometric sequence.Taught in 3.9 Matrix transformations
function
A rule that gives exactly one output for each input in its domain.Taught in 2.2 Function concepts, domain and range
future value
The value after a given number of periods (FVFV); in a calculator's finance solver with no regular payments, it has the opposite sign to the present value, because the money moves the other way.Taught in 1.4 Compound interest and depreciation

G

general form of a line
The equation of a straight line written as ax+by+d=0ax + by + d = 0, with every term on one side and aa and bb not both zero; it can describe any line, including a vertical one.Taught in 2.1 Straight-line equations
general solution
The solution of a differential equation written with arbitrary constants, one for a first-order equation and two for a second-order one, so that it describes a whole family of solution curves.Taught in 5.14 Setting up and separating differential equations
geometric sequence
A sequence with a non-zero first term in which each term after the first is the term before multiplied by the same non-zero number rr, the common ratio, as in 5,15,45,135,…5, 15, 45, 135, \dots where r=3r = 3.Taught in 1.3 Geometric sequences and series
geometric series
The sum of the terms of a geometric sequence; for a common ratio r≠1r \ne 1 the first nn terms add up to Sn=u1(rn−1)r−1S_{n} = \frac{u_{1}(r^{n} - 1)}{r - 1}.Taught in 1.3 Geometric sequences and series
global maximum
The greatest value a function takes anywhere on its domain (the global minimum is the least value); it may be at a local maximum or, when the domain has end points, at an end point.Taught in 5.6 Local maximum and minimum points
gradient
The steepness of a straight line, written mm: the change in yy divided by the change in xx between any two points on it, positive when the line rises from left to right, negative when it falls, and undefined for a vertical line.Taught in 2.1 Straight-line equations
gradient function
Another name for the derivative f′(x)f'(x): the function that gives the gradient of the tangent to the curve y=f(x)y = f(x) at each value of xx.Taught in 5.1 Limits and the derivative as a rate of change
gradient-intercept form
The equation of a non-vertical line written as y=mx+cy = mx + c, where mm is the gradient and the line crosses the yy-axis at (0,c)(0, c).Taught in 2.1 Straight-line equations
graph (graph theory)
A set of vertices joined by edges; only which vertices are joined matters, not where they are drawn or how long the edges look.Taught in 3.14 Graph theory basics

H

half-life
The time a quantity decaying exponentially towards 00 takes to fall to half its value; for N0e−ktN_{0}\mathrm{e}^{-kt} with N0>0N_{0} > 0 and k>0k > 0 it is ln⁡2k\frac{\ln 2}{k}, the same whatever the starting amount.Taught in 2.9 Further function models
Hamiltonian cycle
A cycle that visits every vertex of a graph exactly once and returns to its start vertex.Taught in 3.16 Spanning trees, postman and salesman problems
Hamiltonian path
A path that visits every vertex of a graph exactly once; there is no simple test for whether one exists.Taught in 3.16 Spanning trees, postman and salesman problems
histogram
A chart of grouped data on a continuous scale, with one bar per class and no gaps between bars; when every class has the same width, each bar's height is the class frequency.Taught in 4.2 Data presentation
horizontal stretch
For q>0q > 0, the transformation from y=f(x)y = f(x) to y=f(qx)y = f(qx), with scale factor 1q\frac{1}{q}, not qq: every xx-coordinate is divided by qq, every yy-coordinate stays the same, and the yy-axis is invariant.Taught in 2.8 Transformations of graphs

I

identity matrix
The square matrix II with 11s on the leading diagonal and 00s everywhere else, so that AI=IA=AAI = IA = A for any square matrix AA of the same order.Taught in 1.14 Matrices, determinants and inverses
imaginary part
For z=a+biz = a + b\mathrm{i} with aa and bb real, the imaginary part is the real number Im(z)=b\mathrm{Im}(z) = b, not bib\mathrm{i}.Taught in 1.12 Complex numbers in Cartesian form
increasing function
A function is increasing on an interval if its output gets bigger as xx gets bigger there, so a<ba < b gives f(a)<f(b)f(a) < f(b); this holds whenever f′(x)>0f'(x) > 0 for every xx in the interval.Taught in 5.2 Increasing and decreasing functions
indefinite integral
The family of all antiderivatives of ff, written ∫f(x) dx=F(x)+C\int f(x)\,\mathrm{d}x = F(x) + C; for a power, ∫xn dx=xn+1n+1+C\int x^{n}\,\mathrm{d}x = \frac{x^{n+1}}{n+1} + C only when n≠−1n \neq -1.Taught in 5.5 Antidifferentiation, definite integrals and area
independent events
Events where one happening does not change the probability of the other, so P(A∩B)=P(A) P(B)\mathrm{P}(A \cap B) = \mathrm{P}(A)\,\mathrm{P}(B); two mutually exclusive events that each have a non-zero probability are never independent.Taught in 4.6 Probability diagrams, rules and independence
initial condition
A known value at the start of a situation, usually at t=0t = 0, which you substitute into a model to find a parameter.Taught in 2.6 Fitting and testing models
initial state matrix
The column s0\boldsymbol{s}_0 of probabilities (or proportions) of being in each state at the start, in the same order as the transition matrix TT; its entries add up to 11, and after nn steps the state matrix is sn=Tns0\boldsymbol{s}_n = T^{n}\boldsymbol{s}_0.Taught in 4.19 Transition matrices and Markov chains
integration by inspection
Finding an integral by recognising the integrand as the derivative of a known function, often by reversing the chain rule, then differentiating the guess to check it and fix any constant factor.Taught in 5.11 Standard integrals and substitution
integration by substitution
Finding ∫f(g(x)) g′(x) dx\int f(g(x))\,g'(x)\,\mathrm{d}x by letting u=g(x)u = g(x), so that g′(x) dxg'(x)\,\mathrm{d}x becomes du\mathrm{d}u, integrating with respect to uu and writing the answer in terms of xx again; for a definite integral, you can change the limits to values of uu instead of writing xx back.Taught in 5.11 Standard integrals and substitution
interpolation
Using a model to predict a value inside the range of data it was built from; usually more reliable than extrapolation, but only as good as the model's fit.Taught in 2.6 Fitting and testing models
interquartile range
The difference between the upper and lower quartiles, IQR=Q3−Q1\mathrm{IQR} = Q_{3} - Q_{1}, which measures the spread of the middle half of the data.Taught in 4.1 Sampling, data quality and outliers
invariant
Unchanged by a transformation: an invariant point maps to itself, such as a point on the xx-axis under the vertical stretch y=pf(x)y = pf(x), or on the mirror line under a reflection.Taught in 2.8 Transformations of graphs
inverse function
The function f−1f^{-1} that undoes a one-to-one function ff: if f(a)=bf(a) = b then f−1(b)=af^{-1}(b) = a, its graph is the reflection of the graph of ff in y=xy = x, its domain is the range of ff, and f−1(x)f^{-1}(x) is not 1f(x)\frac{1}{f(x)}.Taught in 2.2 Function concepts, domain and range
inverse matrix
For a square matrix AA with det⁡A≠0\det A \neq 0, the matrix A−1A^{-1} with AA−1=A−1A=IAA^{-1} = A^{-1}A = I; a matrix with det⁡A=0\det A = 0 has no inverse, and Ax=bA\boldsymbol{x} = \boldsymbol{b} gives x=A−1b\boldsymbol{x} = A^{-1}\boldsymbol{b}, multiplying on the left.Taught in 1.14 Matrices, determinants and inverses
inverse normal
Working back from a probability to a value: for a normal variable XX with known mean and standard deviation, finding kk so that P(X<k)=p\mathrm{P}(X < k) = p; for P(X>k)=p\mathrm{P}(X > k) = p, use the area 1−p1 - p or set the GDC's tail to right.Taught in 4.9 Normal distribution
inverse trigonometric function
sin⁡−1\sin^{-1}, cos⁡−1\cos^{-1} and tan⁡−1\tan^{-1} give an angle from a ratio, but a calculator returns only one angle (sin⁡−1\sin^{-1} and tan⁡−1\tan^{-1} between −90∘-90^{\circ} and 90∘90^{\circ}, cos⁡−1\cos^{-1} between 0∘0^{\circ} and 180∘180^{\circ}), so other angles with the same ratio may exist.Taught in 3.2 Triangle trigonometry
inverse variation
A relationship y=axny = ax^{n} with a≠0a \ne 0 and nn a negative integer, such as y=axy = \frac{a}{x}; the yy-axis is a vertical asymptote, and xyxy is constant only when n=−1n = -1.Taught in 2.5 Modelling with functions

K

Kruskal's algorithm
A way to find a minimum spanning tree: take the edges in increasing order of weight, add each one unless it makes a cycle, and stop at n−1n - 1 edges for nn vertices.Taught in 3.16 Spanning trees, postman and salesman problems

L

laws of exponents
The rules for powers of a non-zero base aa with integer exponents: am×an=am+na^{m} \times a^{n} = a^{m+n} and am÷an=am−na^{m} \div a^{n} = a^{m-n} (both need the same base), (am)n=amn(a^{m})^{n} = a^{mn}, a0=1a^{0} = 1 and a−n=1ana^{-n} = \frac{1}{a^{n}}.Taught in 1.5 Exponents and introductory logarithms
laws of logarithms
The rules log⁡axy=log⁡ax+log⁡ay\log_{a} xy = \log_{a} x + \log_{a} y, log⁡axy=log⁡ax−log⁡ay\log_{a} \frac{x}{y} = \log_{a} x - \log_{a} y and log⁡axm=mlog⁡ax\log_{a} x^{m} = m\log_{a} x, which hold for a,x,y>0a, x, y > 0 with a≠1a \neq 1; there is no law for the log of a sum such as log⁡a(x+y)\log_{a}(x + y).Taught in 1.9 Laws of logarithms
least squares regression
Fitting a line or curve of a chosen type (linear, quadratic, cubic, exponential, power or sine) by choosing its parameters so that the sum of the squared vertical distances from the data points to it is as small as possible; for exponential and power models the GDC does this for ln⁡y\ln y rather than yy.Taught in 4.13 Non-linear regression and R²
limit
The value that f(x)f(x) gets closer and closer to as xx gets closer to a number aa from both sides (whatever f(a)f(a) is, or even if it doesn't exist), or as xx grows without bound; in this course you estimate it from a table, a graph or a GDC table.Taught in 5.1 Limits and the derivative as a rate of change
line of best fit
A straight line drawn by eye that follows the trend of the points on a scatter diagram and passes through the mean point (xˉ,yˉ)(\bar{x}, \bar{y}).Taught in 4.4 Correlation and regression (y on x)
linear combination of random variables
A variable such as a1X1+a2X2+⋯+anXna_{1}X_{1} + a_{2}X_{2} + \dots + a_{n}X_{n} built by multiplying random variables by constants and adding; its expected value is always the same combination of the expected values, while its variance is a12Var(X1)+⋯+an2Var(Xn)a_{1}^{2}\mathrm{Var}(X_{1}) + \dots + a_{n}^{2}\mathrm{Var}(X_{n}) only when the variables are independent, with every term added even for a difference.Taught in 4.14 Linear combinations and unbiased estimates
linearizing data
Taking logarithms of one or both variables so that curved data lie on a straight line: for positive data, ln⁡y\ln y against xx is straight for an exponential model y=kaxy = ka^{x} and ln⁡y\ln y against ln⁡x\ln x is straight for a power model y=kxny = kx^{n}.Taught in 2.10 Logarithmic scales and linearizing data
local maximum
A point on a graph that is at least as high as every point near it; it need not give the greatest value on the domain, which may be at an end point.Taught in 2.4 Key features of graphs
local minimum
A point on a graph that is at least as low as every point near it; it need not give the least value on the domain, which may be at an end point.Taught in 2.4 Key features of graphs
log-log graph
A graph with logarithmic scales on both axes, or of ln⁡y\ln y against ln⁡x\ln x, on which data from a power model y=kxny = kx^{n} with k>0k > 0 and x>0x > 0 lie on a straight line with gradient nn.Taught in 2.10 Logarithmic scales and linearizing data
logarithm
For a>0a > 0, a≠1a \neq 1 and b>0b > 0, log⁡ab\log_{a} b is the exponent xx for which ax=ba^{x} = b, so log⁡101000=3\log_{10} 1000 = 3 because 103=100010^{3} = 1000.Taught in 1.5 Exponents and introductory logarithms
logarithmic scale
A scale on which equal distances stand for equal multiplying factors, such as each gridline being 1010 times the one before, used when positive values cover a very wide range.Taught in 2.10 Logarithmic scales and linearizing data
logistic function
A function f(x)=L1+Ce−kxf(x) = \frac{L}{1 + C\mathrm{e}^{-kx}} with L,C,k>0L, C, k > 0, used to model restricted growth: its S-shaped graph rises almost exponentially while f(x)f(x) is small compared with LL, then slows and levels off towards LL.Taught in 2.9 Further function models
lower bound
For a value rounded to the nearest uu, the rounded value minus u2\frac{u}{2}, which the true value can equal (7.37.3 to 11 d.p. has lower bound 7.257.25); in graph theory, a value that the length of the shortest travelling salesman tour cannot be less than.Taught in 1.6 Rounding, bounds and percentage error

M

magnitude of a vector
The length of a vector, ∣v∣=v12+v22+v32\lvert\boldsymbol{v}\rvert = \sqrt{v_{1}^{2} + v_{2}^{2} + v_{3}^{2}} for v=v1i+v2j+v3k\boldsymbol{v} = v_{1}\boldsymbol{i} + v_{2}\boldsymbol{j} + v_{3}\boldsymbol{k}, which is a scalar and never negative.Taught in 3.10 Vectors in 2D and 3D
Markov chain
A process that moves between a fixed set of states in steps, where the probability of each possible next state depends only on the current state, not on the states before it.Taught in 4.19 Transition matrices and Markov chains
matched pairs
Two measurements on each subject, such as before and after, analysed by testing the differences as one sample.Taught in 4.18 Further hypothesis tests and errors
mathematical model
A function or equation that describes a real situation approximately, and is only valid over a domain the context allows.Taught in 2.2 Function concepts, domain and range
matrix
A rectangular array of numbers in rows and columns: a matrix with mm rows and nn columns has order m×nm \times n (rows first), and aija_{ij} is the element in row ii, column jj.Taught in 1.14 Matrices, determinants and inverses
matrix multiplication
Forming the product ABAB, which exists only when AA has as many columns as BB has rows; each element of ABAB is a row of AA times a column of BB (multiply matching pairs, then add), and in general AB≠BAAB \neq BA.Taught in 1.14 Matrices, determinants and inverses
mean
The total of the data values divided by how many values there are, written xˉ\bar{x} (μ\mu for a whole population).Taught in 4.3 Central tendency and dispersion
median
The middle value of the data when it is put in order; with an even number of values it is the mean of the middle two.Taught in 4.2 Data presentation
minimum spanning tree
A spanning tree of a connected weighted graph with the least possible total weight; there may be more than one, all with the same weight.Taught in 3.16 Spanning trees, postman and salesman problems
modal class
For grouped data with equal class widths, the class with the highest frequency.Taught in 4.3 Central tendency and dispersion
modulus of a complex number
For z=a+biz = a + b\mathrm{i} with aa and bb real, ∣z∣=a2+b2\lvert z\rvert = \sqrt{a^{2} + b^{2}}, the distance from 00 to zz on an Argand diagram; it is never negative.Taught in 1.12 Complex numbers in Cartesian form
mutually exclusive events
Events that cannot both happen, so P(A∩B)=0\mathrm{P}(A \cap B) = 0; this is not the same as being independent.Taught in 4.6 Probability diagrams, rules and independence

N

natural logarithm
The logarithm with base e\mathrm{e}, written ln⁡x=log⁡ex\ln x = \log_{\mathrm{e}} x and defined only for x>0x > 0, so ln⁡b=x\ln b = x means ex=b\mathrm{e}^{x} = b.Taught in 1.5 Exponents and introductory logarithms
nearest neighbour algorithm
Building a tour by going from the start to the nearest unvisited vertex each time, then back to the start; its weight is an upper bound for the travelling salesman problem, not always the best tour.Taught in 3.16 Spanning trees, postman and salesman problems
nearest neighbour interpolation
Estimating a quantity at a point as the value at its nearest site, so every point in a Voronoi cell gets the value of that cell's site.Taught in 3.6 Voronoi diagrams
normal distribution
A continuous probability distribution whose curve is symmetric and bell-shaped about its mean μ\mu; X∼N(μ,σ2)X \sim \mathrm{N}(\mu, \sigma^{2}) has mean μ\mu and variance σ2\sigma^{2}, so its standard deviation is σ\sigma.Taught in 4.9 Normal distribution
normal to a curve
The straight line through a point (a,f(a))(a, f(a)) of the curve y=f(x)y = f(x) that is perpendicular to the tangent there; its gradient is −1f′(a)-\frac{1}{f'(a)} when f′(a)≠0f'(a) \neq 0, and it is the vertical line x=ax = a when f′(a)=0f'(a) = 0.Taught in 5.4 Tangents and normals
null hypothesis
The statement H0\mathrm{H}_0 assumed true during a hypothesis test, such as no association, no difference between population means, or data fitting a claimed distribution; a test either rejects it or does not, and never proves it.Taught in 4.11 Chi-squared tests and the t-test

O

odd vertex
A vertex whose degree is odd; every graph has an even number of odd vertices.Taught in 3.16 Spanning trees, postman and salesman problems
one-tailed test
A test whose H1\mathrm{H}_1 states a direction, such as μ1>μ2\mu_1 > \mu_2, so the p-value is the area in one tail only.Taught in 4.11 Chi-squared tests and the t-test
one-to-one function
A function in which no two different inputs give the same output, so every horizontal line meets its graph at most once and the function has an inverse.Taught in 2.2 Function concepts, domain and range
optimisation
Finding the greatest or least value of a quantity in context over the values its variable can take, usually by comparing the points where the derivative is zero with the ends of the domain.Taught in 5.7 Optimisation problems
outlier
A data value less than Q1−1.5×IQRQ_{1} - 1.5 \times \mathrm{IQR} or greater than Q3+1.5×IQRQ_{3} + 1.5 \times \mathrm{IQR}; some outliers are valid data and some are errors, so check the context before removing one.Taught in 4.1 Sampling, data quality and outliers

P

p-value
The probability, assuming H0\mathrm{H}_0 is true, of a result at least as extreme as the one observed; reject H0\mathrm{H}_0 when it is less than the significance level. It is not the probability that H0\mathrm{H}_0 is true.Taught in 4.11 Chi-squared tests and the t-test
parabola
The U-shaped graph of a quadratic function y=ax2+bx+cy = ax^{2} + bx + c, which opens upwards when a>0a > 0 and downwards when a<0a < 0.Taught in 2.5 Modelling with functions
parallel forms reliability
A reliability check in which the same group takes two equivalent versions of a test; a strong positive correlation between the two sets of scores suggests the test is reliable.Taught in 4.12 Data collection, reliability and validity
parameter
A constant in a model, such as kk, aa or cc in f(x)=kax+cf(x) = ka^{x} + c, found from the situation and then kept fixed while the variable changes.Taught in 2.6 Fitting and testing models
parametric form
A line written as one equation per coordinate in terms of a parameter λ\lambda, such as x=x0+λlx = x_0 + \lambda l, y=y0+λmy = y_0 + \lambda m, z=z0+λnz = z_0 + \lambda n, where (x0,y0,z0)(x_0, y_0, z_0) is a point on the line and (lmn)\begin{pmatrix}l\\m\\n\end{pmatrix} is a direction vector; each value of λ\lambda gives one point.Taught in 3.11 Vector equations of lines
particular solution
The one solution of a differential equation that also fits the given conditions, such as a known value at t=0t = 0; the conditions fix the constants in the general solution, one condition for each constant.Taught in 5.14 Setting up and separating differential equations
path
A walk in which no vertex is visited more than once, so no edge is repeated either.Taught in 3.16 Spanning trees, postman and salesman problems
Pearson's product-moment correlation coefficient
The number rr, with −1≤r≤1-1 \le r \le 1, that measures the strength and direction of linear correlation only; rr near 00 means no straight-line link, though the variables may still follow a curve.Taught in 4.4 Correlation and regression (y on x)
percentage error
How far an approximate value vAv_{A} is from the exact value vEv_{E}, as a percentage of the exact value: ∣vA−vEvE∣×100%\left|\frac{v_{A} - v_{E}}{v_{E}}\right| \times 100\%; always divide by the exact value, never by the approximation.Taught in 1.6 Rounding, bounds and percentage error
percentile
The kkth percentile is the value with k%k\% of the data below it, so the median is the 50th percentile.Taught in 4.2 Data presentation
period
The horizontal length of one complete cycle of a repeating graph; for asin⁡(bx)+da\sin(bx) + d or acos⁡(bx)+da\cos(bx) + d with b>0b > 0 it is 360∘b\frac{360^{\circ}}{b} in degrees, or 2πb\frac{2\pi}{b} in radians.Taught in 2.5 Modelling with functions
perpendicular bisector
The line through the midpoint of a line segment at right angles to it; in a plane, a point lies on it exactly when it is the same distance from both ends of the segment.Taught in 3.5 Perpendicular bisectors
perpendicular lines
Lines that meet at a right angle; two non-vertical lines with gradients m1m_{1} and m2m_{2} are perpendicular exactly when m1×m2=−1m_{1} \times m_{2} = -1, and a horizontal line is perpendicular to a vertical one.Taught in 2.1 Straight-line equations
phase portrait
A diagram in the (x,y)(x, y) plane showing the trajectories of a pair of coupled differential equations, with arrows for the direction of motion as time increases; time itself is not an axis.Taught in 5.17 Phase portraits for coupled equations
phase shift
The horizontal translation of a sinusoidal graph: in f(x)=asin⁡(b(x−c))+df(x) = a\sin(b(x - c)) + d or acos⁡(b(x−c))+da\cos(b(x - c)) + d it is cc, to the right when c>0c > 0, so acos⁡(bx+B)a\cos(bx + B) with b>0b > 0 is shifted Bb\frac{B}{b} to the left, not BB.Taught in 1.13 Complex numbers in polar and exponential form
piecewise function
A function that uses different rules on different intervals of its domain, with each input using exactly one rule; check which interval an input is in, endpoints included, before substituting.Taught in 2.5 Modelling with functions
point of inflexion
A point on a graph where the concavity changes, from concave-up to concave-down or the other way; if f′′f'' exists there it equals 00, but f′′(x)=0f''(x) = 0 alone does not make a point of inflexion, and the gradient there need not be 00.Taught in 5.10 Second derivative and concavity
point-gradient form
The equation y−y1=m(x−x1)y - y_{1} = m(x - x_{1}) of the straight line with gradient mm that passes through the point (x1,y1)(x_{1}, y_{1}).Taught in 2.1 Straight-line equations
Poisson distribution
The distribution X∼Po(m)X \sim \mathrm{Po}(m) of the number of events in a fixed interval of time or space, when the events happen independently at a uniform average rate of mm per interval; XX takes the values 0,1,2,…0, 1, 2, \ldots and has mean and variance both equal to mm.Taught in 4.17 Poisson distribution
polar form
A non-zero complex number written z=r(cos⁡θ+isin⁡θ)=r cis θz = r(\cos\theta + \mathrm{i}\sin\theta) = r\,\mathrm{cis}\,\theta, where r=∣z∣>0r = \lvert z\rvert > 0 is its modulus and θ=arg⁡z\theta = \arg z is its argument; also called modulus-argument form.Taught in 1.13 Complex numbers in polar and exponential form
polynomial
A sum of terms axnax^{n}, where each coefficient aa is a constant and each power nn is a whole number 0,1,2,…0, 1, 2, \ldots; the highest power with a non-zero coefficient is its degree.Taught in 1.8 Solving equations with technology
population
The whole group of people or things that a statistical question is about.Taught in 4.1 Sampling, data quality and outliers
population product-moment correlation coefficient
The correlation ρ\rho between two variables in the whole population, estimated by the sample's rr; testing H0\mathrm{H}_0: ρ=0\rho = 0 assumes the data come from a bivariate normal distribution.Taught in 4.18 Further hypothesis tests and errors
position vector
The arrow from a fixed origin O to a point A, labelled OA→\overrightarrow{OA} or a\boldsymbol{a}; its components equal the coordinates of A.Taught in 3.10 Vectors in 2D and 3D
power rule
The rule that the derivative of axnax^{n} is anxn−1anx^{n-1} (multiply by the power, then lower the power by one), used at SL for integer nn and at HL also for rational nn; it does not apply to an exponential such as axa^{x}, where the variable is in the power.Taught in 5.3 Differentiation of integer-power polynomials
predator–prey model
A pair of coupled differential equations for a prey population and a predator population, in which meetings between them lower the prey's growth rate and raise the predators' growth rate.Taught in 5.16 Euler's method
present value
The value at the start (PVPV), such as the amount invested or borrowed; in a calculator's finance solver with no regular payments, it has the opposite sign to the future value, because the money moves the other way.Taught in 1.4 Compound interest and depreciation
Prim's algorithm
A way to find a minimum spanning tree: from a start vertex, keep adding the least-weight edge from a vertex in the tree to a vertex not yet in it; it can be run directly on a table (the matrix method).Taught in 3.16 Spanning trees, postman and salesman problems
principal axis
The horizontal line y=dy = d halfway between the maximum and minimum values of asin⁡(bx)+da\sin(bx) + d or acos⁡(bx)+da\cos(bx) + d.Taught in 2.5 Modelling with functions
probability distribution
Every value a discrete random variable XX can take, together with its probability P(X=x)\mathrm{P}(X = x), given as a table or a formula. Each probability lies between 00 and 11, and together they add up to 11.Taught in 4.7 Discrete random variables and expected value
product rule
The rule for differentiating a product of two functions: if y=uvy = uv, then dydx=udvdx+vdudx\frac{\mathrm{d}y}{\mathrm{d}x} = u\frac{\mathrm{d}v}{\mathrm{d}x} + v\frac{\mathrm{d}u}{\mathrm{d}x}.Taught in 5.9 Differentiation rules and related rates
Pythagorean identity
The identity cos⁡2θ+sin⁡2θ=1\cos^{2}\theta + \sin^{2}\theta = 1, true for every angle θ\theta, which is Pythagoras' theorem for a point on the unit circle.Taught in 3.8 Unit-circle trigonometry and equations

Q

quadratic function
A function of the form f(x)=ax2+bx+cf(x) = ax^{2} + bx + c, where aa, bb and cc are constants and a≠0a \ne 0; its graph is a parabola.Taught in 2.5 Modelling with functions
quartile
One of the values that split ordered data into quarters: about a quarter of the data lies below the lower quartile Q1Q_{1} and about three quarters below the upper quartile Q3Q_{3}; methods differ, so a GDC and a hand method can give slightly different values.Taught in 4.1 Sampling, data quality and outliers
quota sampling
Fixing how many members to take from each group, then letting the collector choose anyone who fits until each quota is filled; it is not random, so it can be biased.Taught in 4.1 Sampling, data quality and outliers
quotient rule
The rule for differentiating one function divided by another: if y=uvy = \frac{u}{v}, then dydx=vdudx−udvdxv2\frac{\mathrm{d}y}{\mathrm{d}x} = \frac{v\frac{\mathrm{d}u}{\mathrm{d}x} - u\frac{\mathrm{d}v}{\mathrm{d}x}}{v^{2}}, with the two terms on top in that order.Taught in 5.9 Differentiation rules and related rates

R

radian
A unit of angle: one radian is the angle at the centre of a circle when the arc it cuts off is as long as the radius, so a full turn is 2π2\pi radians and π\pi radians =180∘= 180^{\circ}.Taught in 3.7 Radians, arc length and sector area
random sample
A sample chosen so that every member of the population has the same chance of being selected.Taught in 4.1 Sampling, data quality and outliers
random variable
A quantity whose value is decided by the outcome of a random trial, such as the number of cars a randomly chosen household owns. It is named with a capital letter, such as XX; a lower-case xx is one value it can take.Taught in 4.7 Discrete random variables and expected value
range
The set of output values a function takes over its domain; this is not the range of a data set, which is the largest value minus the smallest.Taught in 2.2 Function concepts, domain and range
range of a data set
The largest data value minus the smallest, a single number that measures spread; it is not the range of a function.Taught in 4.2 Data presentation
rate of change
How fast one quantity changes for each unit change in another at a given point, given by a derivative such as dVdr\frac{\mathrm{d}V}{\mathrm{d}r} or dsdt\frac{\mathrm{d}s}{\mathrm{d}t}; its units are units of the first quantity per unit of the second, and it is negative when the first quantity is decreasing.Taught in 5.1 Limits and the derivative as a rate of change
rational exponent
An exponent that is a fraction: for a>0a > 0, an integer mm and a positive integer nn, amn=(an)ma^{\frac{m}{n}} = \left(\sqrt[n]{a}\right)^{m}, so the denominator gives the root and the numerator the power.Taught in 1.10 Rational exponents
real part
For z=a+biz = a + b\mathrm{i} with aa and bb real, the real part is Re(z)=a\mathrm{Re}(z) = a.Taught in 1.12 Complex numbers in Cartesian form
real value
What a future amount of money is worth in today's prices once inflation, the yearly percentage rise in prices, is allowed for; it is less than the future value whenever inflation is positive.Taught in 1.4 Compound interest and depreciation
reflection
A transformation that flips a graph over a mirror line, sending each point to the point the same distance away on the other side; y=−f(x)y = -f(x) is the reflection of y=f(x)y = f(x) in the xx-axis and y=f(−x)y = f(-x) is its reflection in the yy-axis.Taught in 2.8 Transformations of graphs
regression line
The least-squares line of yy on xx, y=ax+by = ax + b, used to predict yy from a value of xx inside the range of the data when the correlation is strong; it cannot reliably predict xx from yy.Taught in 4.4 Correlation and regression (y on x)
regular Markov chain
A Markov chain whose transition matrix TT has some power TkT^{k} with every entry strictly positive; such a chain has exactly one steady state, and the state matrix approaches it from any initial state.Taught in 4.19 Transition matrices and Markov chains
related rates
Rates of change over time of two or more quantities linked by an equation, so that one rate can be found from another with the chain rule, for example dAdt=dAdr×drdt\frac{\mathrm{d}A}{\mathrm{d}t} = \frac{\mathrm{d}A}{\mathrm{d}r} \times \frac{\mathrm{d}r}{\mathrm{d}t}.Taught in 5.9 Differentiation rules and related rates
relative frequency
The number of times an event happened divided by the number of trials: an experimental estimate of its probability, unlike the theoretical n(A)n(U)\frac{n(A)}{n(U)}, which needs equally likely outcomes.Taught in 4.5 Probability foundations and expected value
relative position
The position of B relative to A is the vector from A to B, AB→=b−a\overrightarrow{AB} = \boldsymbol{b} - \boldsymbol{a} (B's position minus A's, not the other way round); for two moving objects it is rB(t)−rA(t)\boldsymbol{r}_{B}(t) - \boldsymbol{r}_{A}(t) at the same time tt, and its magnitude is the distance between them.Taught in 3.12 Kinematics with vectors
reliability
How consistently a method gives the same results when it is repeated under the same conditions; a reliable method is not necessarily valid, because it can consistently measure the wrong thing.Taught in 4.12 Data collection, reliability and validity
residual
For one data point, the observed yy-value minus the value the model predicts at the same xx; it is positive when the point lies above the curve and negative when it lies below.Taught in 4.13 Non-linear regression and R²
root
A value of the unknown that makes an equation true; the roots of f(x)=0f(x) = 0 are the zeros of ff.Taught in 1.8 Solving equations with technology

S

saddle point
An equilibrium point of a linear system whose eigenvalues are real with opposite signs: trajectories come in alongside the negative eigenvalue's eigenvector line and leave alongside the positive one's, so it is unstable.Taught in 5.17 Phase portraits for coupled equations
sample
The part of a population that data is actually collected from, used to draw conclusions about the whole population.Taught in 4.1 Sampling, data quality and outliers
sample space
The set of all possible outcomes of a trial, written UU; n(U)n(U) is the number of outcomes in it.Taught in 4.5 Probability foundations and expected value
sampling bias
A built-in tendency of a sampling method to include some kinds of member of the population more often than others, so the sample misrepresents the population; a bigger sample chosen the same way does not remove it.Taught in 4.1 Sampling, data quality and outliers
scalar product
The number v⋅w=v1w1+v2w2+v3w3\boldsymbol{v} \cdot \boldsymbol{w} = v_{1}w_{1} + v_{2}w_{2} + v_{3}w_{3}, found by multiplying matching components and adding, which also equals ∣v∣∣w∣cos⁡θ\lvert\boldsymbol{v}\rvert\lvert\boldsymbol{w}\rvert\cos\theta where θ\theta is the angle between the vectors; for non-zero vectors it is 00 exactly when they are perpendicular.Taught in 3.13 Scalar and vector products
scatter diagram
A graph that plots each pair (x,y)(x, y) of bivariate data as one point, to show how the two variables are related.Taught in 4.4 Correlation and regression (y on x)
second derivative
The derivative of the derivative, written f′′(x)f''(x) or d2ydx2\frac{\mathrm{d}^{2}y}{\mathrm{d}x^{2}}; it measures how fast the gradient is changing.Taught in 5.10 Second derivative and concavity
second derivative test
A way to classify a stationary point at x=ax = a: if f′′(a)>0f''(a) > 0 it is a local minimum, if f′′(a)<0f''(a) < 0 it is a local maximum, and if f′′(a)=0f''(a) = 0 the test decides nothing.Taught in 5.10 Second derivative and concavity
second-order differential equation
A differential equation whose highest derivative is a second derivative, such as d2xdt2=f(x,dxdt,t)\frac{\mathrm{d}^{2}x}{\mathrm{d}t^{2}} = f\left(x, \frac{\mathrm{d}x}{\mathrm{d}t}, t\right); putting y=dxdty = \frac{\mathrm{d}x}{\mathrm{d}t} turns it into the coupled first-order equations dxdt=y\frac{\mathrm{d}x}{\mathrm{d}t} = y and dydt=f(x,y,t)\frac{\mathrm{d}y}{\mathrm{d}t} = f(x, y, t).Taught in 5.18 Second-order differential equations
sector
The region of a circle between two radii and the arc joining their ends; with radius rr and angle θ\theta degrees at the centre, its area is θ360×πr2\frac{\theta}{360} \times \pi r^{2}.Taught in 3.4 Arc length and sector area
semi-log graph
A graph with a logarithmic scale on one axis only, usually the vertical one, on which data from an exponential model y=kaxy = ka^{x} with k,a>0k, a > 0 lie on a straight line.Taught in 2.10 Logarithmic scales and linearizing data
separation of variables
A method for solving dydx=f(x) g(y)\frac{\mathrm{d}y}{\mathrm{d}x} = f(x)\,g(y) where g(y)≠0g(y) \neq 0: rearrange to ∫1g(y) dy=∫f(x) dx\int \frac{1}{g(y)}\,\mathrm{d}y = \int f(x)\,\mathrm{d}x, then integrate both sides with one constant.Taught in 5.14 Setting up and separating differential equations
sequence
An ordered list of numbers, called terms; unu_{n} is the nnth term itself, not the sum of the terms up to it.Taught in 1.2 Arithmetic sequences and series
series
The sum of the terms of a sequence; SnS_{n} is the sum of the first nn terms, not the nnth term.Taught in 1.2 Arithmetic sequences and series
sigma notation
A short way to write a sum: ∑r=abf(r)\sum_{r=a}^{b} f(r) means put every whole number rr from aa to bb, both included, into f(r)f(r) and add the results.Taught in 1.2 Arithmetic sequences and series
significance level
The probability threshold, such as 5%5\%, chosen before a hypothesis test: H0\mathrm{H}_0 is rejected when the p-value is less than it.Taught in 4.11 Chi-squared tests and the t-test
significant figures
The digits of a number counted from its first non-zero digit, so 0.04070.0407 has 33 significant figures; zeros between non-zero digits count, and so does a zero kept at the end after rounding (5.705.70), but zeros at the end of a whole number may only hold places (86 00086\,000 to 22 s.f.); IB answers are exact or given to 33 significant figures unless the question says otherwise.Taught in 1.6 Rounding, bounds and percentage error
simple graph
A graph with no loops and no multiple edges, so each pair of vertices is joined by at most one edge.Taught in 3.14 Graph theory basics
simple interest
Interest worked out each period on the original amount only, so the same sum is added every period and the balances form an arithmetic sequence rather than a geometric one.Taught in 1.2 Arithmetic sequences and series
simple random sampling
Choosing a sample so that every possible group of the required size is equally likely, for example by picking numbers at random from a full list of the population.Taught in 4.1 Sampling, data quality and outliers
sine rule
In any triangle ABCABC, asin⁡A=bsin⁡B=csin⁡C\frac{a}{\sin A} = \frac{b}{\sin B} = \frac{c}{\sin C}, where each side is opposite the angle with the same letter.Taught in 3.2 Triangle trigonometry
sinusoidal function
A wave-shaped function asin⁡(bx)+da\sin(bx) + d or acos⁡(bx)+da\cos(bx) + d with b>0b > 0; its period is 360∘b\frac{360^{\circ}}{b} when bxbx is in degrees and 2πb\frac{2\pi}{b} when it is in radians.Taught in 2.5 Modelling with functions
site
One of the given points a Voronoi diagram is built from, such as a shop or a weather station; each site has its own cell.Taught in 3.6 Voronoi diagrams
slant height
On a right cone, the slant height ll is the distance from the vertex to the edge of the base, measured along the sloping surface; l=r2+h2l = \sqrt{r^{2} + h^{2}}, so it is longer than the vertical height hh.Taught in 3.1 3D distance, midpoint and measurement
slope field
A diagram for a differential equation dydx=f(x,y)\frac{\mathrm{d}y}{\mathrm{d}x} = f(x, y) with a short line segment at each grid point (a,b)(a, b), drawn with gradient f(a,b)f(a, b); every solution curve is tangent to the segments it passes through.Taught in 5.15 Slope fields
spanning tree
A subgraph of a connected graph that is a tree and includes every vertex of the graph.Taught in 3.16 Spanning trees, postman and salesman problems
Spearman's rank correlation coefficient
The number rsr_s, with −1≤rs≤1-1 \le r_s \le 1, found by calculating Pearson's rr on the ranks of the data (tied values each take the mean of the ranks they share); it measures how strongly one variable always rises, or always falls, as the other rises, straight-line or curved, and is less affected by outliers than Pearson's rr.Taught in 4.10 Spearman's rank correlation
speed
How fast an object moves, whatever its direction: the magnitude ∣v∣\lvert\boldsymbol{v}\rvert of its velocity, a scalar that is never negative.Taught in 3.10 Vectors in 2D and 3D
stable equilibrium
An equilibrium point that all nearby solutions move towards as time increases; for dxdt=ax+by\frac{\mathrm{d}x}{\mathrm{d}t} = ax + by, dydt=cx+dy\frac{\mathrm{d}y}{\mathrm{d}t} = cx + dy the origin is stable when both eigenvalues are negative or complex with negative real part.Taught in 5.17 Phase portraits for coupled equations
standard deviation
The square root of the variance, measuring how far data values typically lie from the mean; adding a constant to every value leaves it unchanged, and multiplying every value by kk multiplies it by ∣k∣\lvert k\rvert.Taught in 4.3 Central tendency and dispersion
standard form
A way of writing a number as a×10ka \times 10^{k}, where 1≤a<101 \le a < 10 and kk is an integer, such as 3.4×10213.4 \times 10^{21}; calculator notation such as 3.4E21 is not accepted.Taught in 1.1 Standard form
stationary point
A point on a curve where f′(x)=0f'(x) = 0, so the tangent is horizontal; it may be a local maximum, a local minimum or neither.Taught in 5.2 Increasing and decreasing functions
steady state
A column s\boldsymbol{s} of non-negative entries adding up to 11 with Ts=sT\boldsymbol{s} = \boldsymbol{s} for the transition matrix TT: an eigenvector of TT for the eigenvalue 11, scaled so its entries add up to 11; for a regular Markov chain it is unique and gives the long-term probabilities.Taught in 4.19 Transition matrices and Markov chains
stratified sampling
Splitting the population into groups that share a feature, then taking a random sample from each group in proportion to its size, with the rounded group sizes adjusted so they add up to the total sample size.Taught in 4.1 Sampling, data quality and outliers
strongly connected graph
A directed graph in which you can travel from every vertex to every other vertex following the directions of the edges.Taught in 3.14 Graph theory basics
subgraph
A graph made from some of the vertices and edges of another graph, in which every chosen edge keeps both of its end vertices.Taught in 3.14 Graph theory basics
sum of square residuals
SSresSS_{\text{res}}, the total of the squared residuals of a model; for models fitted to the same data a smaller value means a closer fit, and 00 means every point lies on the curve.Taught in 4.13 Non-linear regression and R²
sum to infinity
The value S∞S_{\infty} that the partial sums SnS_{n} of an infinite geometric series get closer and closer to as nn grows; it exists only when the common ratio satisfies ∣r∣<1\lvert r\rvert < 1, and then S∞=u11−rS_{\infty} = \frac{u_{1}}{1 - r}.Taught in 1.11 Infinite geometric series
system of linear equations
Two or more linear equations in the same unknowns, solved together, where a solution makes every equation true at once; in IB AI exams a system has up to 33 unknowns and always exactly one solution.Taught in 1.8 Solving equations with technology
systematic sampling
Choosing every kkth member of an ordered list of the population, where k=Nnk = \frac{N}{n} (rounded down if not whole) for a population of NN and a sample of nn, starting from a member picked at random among the first kk and stopping once there are nn. If nn doesn't divide NN, the last members of the list can never be picked.Taught in 4.1 Sampling, data quality and outliers

T

t-distribution
A symmetric, bell-shaped distribution with heavier tails than the standard normal, fixed by its degrees of freedom (n−1n - 1 for one sample of size nn); for the mean of a normal population it replaces the normal distribution whenever σ\sigma is unknown and estimated by sn−1s_{n-1}, whatever the sample size.Taught in 4.16 Confidence intervals for the mean
t-test
A test about population means using the t-distribution; at SL it is the pooled two-sample t-test, for unpaired samples from normal populations with unknown but equal variances.Taught in 4.11 Chi-squared tests and the t-test
table of least distances
A table giving the weight of the shortest route, direct or not, between every pair of vertices; it turns a practical travelling salesman problem into the classical one on a complete graph.Taught in 3.16 Spanning trees, postman and salesman problems
tangent
The straight line that touches the curve y=f(x)y = f(x) at the point (a,f(a))(a, f(a)) and has the curve's gradient f′(a)f'(a) there, so its equation is y−f(a)=f′(a)(x−a)y - f(a) = f'(a)(x - a).Taught in 5.4 Tangents and normals
test–retest reliability
A reliability check in which the same group takes the same test at two different times; a strong positive correlation between the two sets of scores suggests the test is reliable.Taught in 4.12 Data collection, reliability and validity
toxic waste dump problem
Finding the point in a region whose distance to the nearest site is as large as possible; in IB exams the answer is always at a Voronoi vertex where three edges meet.Taught in 3.6 Voronoi diagrams
trail
A walk in which no edge is used more than once, although vertices may be repeated.Taught in 3.16 Spanning trees, postman and salesman problems
trajectory
The path a solution (x(t),y(t))(x(t), y(t)) of a pair of coupled differential equations traces in the (x,y)(x, y) plane as tt increases.Taught in 5.17 Phase portraits for coupled equations
transformation matrix
A 2×22 \times 2 matrix AA that moves each point x\boldsymbol{x} of the plane to its image AxA\boldsymbol{x} (plus a translation vector b\boldsymbol{b} if there is one), such as a stretch, an enlargement, a rotation about the origin or a reflection in a line through the origin; ∣det⁡A∣\lvert\det A\rvert is its area scale factor.Taught in 3.9 Matrix transformations
transition diagram
A diagram of a Markov chain with one node per state and an arrow, labelled with its probability, for each possible one-step move (a loop for staying in the same state); the probabilities on the arrows leaving any one state add up to 11.Taught in 4.19 Transition matrices and Markov chains
transition matrix
A square matrix of the probabilities of moving between states in one step; in IB notation TijT_{ij} is the probability of moving from state jj to state ii, so each column adds up to 11.Taught in 3.15 Adjacency matrices and walks
translation
A move that slides every point of a graph the same distance in the same direction; the translation by the vector (ab)\begin{pmatrix} a \\ b \end{pmatrix} turns y=f(x)y = f(x) into y=f(x−a)+by = f(x - a) + b, moving it aa units right and bb units up.Taught in 2.8 Transformations of graphs
trapezoidal rule
A way to estimate ∫aby dx\int_{a}^{b} y\,\mathrm{d}x by splitting the interval into nn strips of equal width h=b−anh = \frac{b-a}{n} and treating each strip as a trapezoid, giving 12h((y0+yn)+2(y1+⋯+yn−1))\frac{1}{2}h\left((y_{0} + y_{n}) + 2(y_{1} + \dots + y_{n-1})\right); it overestimates when the curve is concave up across the whole interval and underestimates when it is concave down across the whole interval.Taught in 5.8 Trapezoidal rule
travelling salesman problem
Finding the Hamiltonian cycle of least total weight in a weighted complete graph; a practical problem is first turned into one with a table of least distances.Taught in 3.16 Spanning trees, postman and salesman problems
tree
A connected graph with no cycles; a tree with nn vertices has n−1n - 1 edges.Taught in 3.14 Graph theory basics
tree diagram
A diagram with one set of branches for each stage of an experiment, where the branches from each point add up to 11; you multiply along a path and add across paths, and a later branch can carry a probability that depends on the earlier one.Taught in 4.6 Probability diagrams, rules and independence
trigonometric ratios
For an acute angle θ\theta in a right-angled triangle, sin⁡θ\sin\theta is opposite over hypotenuse, cos⁡θ\cos\theta is adjacent over hypotenuse and tan⁡θ\tan\theta is opposite over adjacent, and each depends only on θ\theta.Taught in 3.2 Triangle trigonometry
two-tailed test
A test whose H1\mathrm{H}_1 uses ≠\neq, such as μ1≠μ2\mu_1 \neq \mu_2, so the p-value counts results as extreme in both tails.Taught in 4.11 Chi-squared tests and the t-test
Type I error
Rejecting the null hypothesis H0\mathrm{H}_0 when H0\mathrm{H}_0 is actually true: a false alarm.Taught in 4.18 Further hypothesis tests and errors
Type II error
Not rejecting the null hypothesis H0\mathrm{H}_0 when H0\mathrm{H}_0 is actually false; its probability can only be found for a stated true value of the parameter.Taught in 4.18 Further hypothesis tests and errors

U

unbiased estimate
An estimate of a population parameter whose expected value over all random samples of the same size equals that parameter, such as xˉ\bar{x} for μ\mu and sn−12=nn−1sn2s_{n-1}^{2} = \frac{n}{n-1}s_{n}^{2} (dividing by n−1n - 1, not nn) for σ2\sigma^{2}.Taught in 4.14 Linear combinations and unbiased estimates
union of events
The event A∪BA \cup B, read "AA or BB": at least one of AA and BB happens, including when both happen.Taught in 4.6 Probability diagrams, rules and independence
unit circle
The circle of radius 11 centred at the origin; the point on it at angle θ\theta anticlockwise from the positive xx-axis is (cos⁡θ,sin⁡θ)(\cos\theta, \sin\theta), which defines sine and cosine for any angle.Taught in 3.8 Unit-circle trigonometry and equations
unit vector
A vector of magnitude 1; the unit vector in the direction of a non-zero vector v\boldsymbol{v} is v∣v∣\frac{\boldsymbol{v}}{\lvert\boldsymbol{v}\rvert}, found by dividing every component by ∣v∣\lvert\boldsymbol{v}\rvert (normalising).Taught in 3.10 Vectors in 2D and 3D
upper bound
For a value rounded to the nearest uu, the rounded value plus u2\frac{u}{2}, which the true value is strictly less than (7.37.3 to 11 d.p. has upper bound 7.357.35); in graph theory, a value that the length of the shortest travelling salesman tour cannot be greater than.Taught in 1.6 Rounding, bounds and percentage error

V

validity
Whether a method measures what it claims to measure, so that its results and conclusions are accurate.Taught in 4.12 Data collection, reliability and validity
variance
The mean of the squared distances of the data values from the mean, equal to the square of the standard deviation; for a data set it divides by nn, the number of values.Taught in 4.3 Central tendency and dispersion
variance of a random variable
The expected squared distance of a random variable from its mean, Var(X)=E((X−μ)2)\mathrm{Var}(X) = \mathrm{E}\left((X - \mu)^{2}\right), in squared units; for constants aa and bb, Var(aX+b)=a2Var(X)\mathrm{Var}(aX + b) = a^{2}\mathrm{Var}(X), so adding a constant does not change it.Taught in 4.14 Linear combinations and unbiased estimates
vector
A quantity with both a magnitude (size) and a direction, such as a displacement, velocity or force, printed in bold as v\boldsymbol{v} or written AB→\overrightarrow{AB}, unlike a scalar, which has a magnitude only.Taught in 3.10 Vectors in 2D and 3D
vector equation of a line
An equation r=a+λb\boldsymbol{r} = \boldsymbol{a} + \lambda\boldsymbol{b}, λ∈R\lambda \in \mathbb{R}, where a\boldsymbol{a} is the position vector of one point on the line and b\boldsymbol{b} is a direction vector of the line, so each value of λ\lambda gives the position vector r\boldsymbol{r} of one point on it.Taught in 3.11 Vector equations of lines
vector product
The vector v×w=∣v∣∣w∣sin⁡θ n\boldsymbol{v} \times \boldsymbol{w} = \lvert\boldsymbol{v}\rvert\lvert\boldsymbol{w}\rvert\sin\theta\,\boldsymbol{n}, where θ\theta is the angle between v\boldsymbol{v} and w\boldsymbol{w} and n\boldsymbol{n} is the unit vector perpendicular to both given by the right-hand screw rule (curl the fingers of your right hand from v\boldsymbol{v} towards w\boldsymbol{w} and your thumb points along n\boldsymbol{n}); its length ∣v×w∣\lvert\boldsymbol{v} \times \boldsymbol{w}\rvert is the area of the parallelogram with adjacent sides v\boldsymbol{v} and w\boldsymbol{w}, and w×v=−(v×w)\boldsymbol{w} \times \boldsymbol{v} = -(\boldsymbol{v} \times \boldsymbol{w}).Taught in 3.13 Scalar and vector products
velocity
The rate of change of position, a vector that gives both how fast and in which direction an object moves; its magnitude is the speed, and in one dimension its sign gives the direction.Taught in 3.10 Vectors in 2D and 3D
Venn diagram
A picture of a sample space as a rectangle, with each event drawn as a circle inside it; where circles overlap, the outcomes belong to both events.Taught in 4.6 Probability diagrams, rules and independence
vertex (graph theory)
One of the points of a graph, also called a node, which edges join to other vertices.Taught in 3.14 Graph theory basics
vertex (Voronoi diagram)
A point where three or more edges meet; it is the same distance from the sites of all the cells that meet there, and no site is closer to it.Taught in 3.6 Voronoi diagrams
vertex of a parabola
The turning point of a parabola, which lies on its axis of symmetry: a minimum point when the parabola opens upwards (a>0a > 0 in y=ax2+bx+cy = ax^{2} + bx + c) and a maximum point when it opens downwards (a<0a < 0).Taught in 2.4 Key features of graphs
vertical stretch
For p>0p > 0, the transformation from y=f(x)y = f(x) to y=pf(x)y = pf(x), with scale factor pp: every yy-coordinate is multiplied by pp, every xx-coordinate stays the same, and the xx-axis is invariant.Taught in 2.8 Transformations of graphs
volume of revolution
The volume of the solid swept out when a region is rotated one full turn (2π2\pi radians or 360∘360^{\circ}) about an axis: for the region between a curve and the xx-axis rotated about the xx-axis it is ∫abπy2 dx\int_{a}^{b} \pi y^{2}\,\mathrm{d}x with xx-values as limits, and for the region between a curve and the yy-axis rotated about the yy-axis it is ∫abπx2 dy\int_{a}^{b} \pi x^{2}\,\mathrm{d}y with yy-values as limits.Taught in 5.12 Areas and volumes of revolution
Voronoi diagram
A division of a region into cells, one for each site, where each cell holds the points that are closer to its site than to any other site.Taught in 3.6 Voronoi diagrams

W

walk
A sequence of edges in which each edge starts at the vertex where the previous edge ended; vertices and edges may be repeated, and the length of the walk is the number of edges it uses.Taught in 3.15 Adjacency matrices and walks
weighted adjacency table
A table with a row and a column for each vertex of a weighted graph, giving the weight (such as a cost, distance or time) of the direct edge from the row's vertex to the column's vertex, with a blank where there is no direct edge.Taught in 3.15 Adjacency matrices and walks
weighted graph
A graph with a number, called a weight, on each edge, standing for something such as a distance, time or cost.Taught in 3.14 Graph theory basics
without replacement
Choosing items one after another without putting each one back, so the total, and the probabilities for the next pick, change after every pick.Taught in 4.6 Probability diagrams, rules and independence

X

x-intercept
A point where a graph meets the xx-axis, crossing or just touching it, found by putting y=0y = 0.Taught in 2.1 Straight-line equations

Y

y-intercept
The point where a graph meets the yy-axis, found by putting x=0x = 0; for the line y=mx+cy = mx + c it is (0,c)(0, c).Taught in 2.1 Straight-line equations

Z

zero of a function
A value aa with f(a)=0f(a) = 0: the xx-coordinate of a point where the graph of y=f(x)y = f(x) meets the xx-axis. The yy-intercept is a zero only if the graph passes through the origin.Taught in 1.8 Solving equations with technology

Χ

χ² goodness of fit test
A test of H0\mathrm{H}_0: observed frequencies fit a claimed distribution; with nn categories and no parameters estimated from the data it has n−1n - 1 degrees of freedom, uses the upper tail only and needs every expected frequency greater than 55.Taught in 4.11 Chi-squared tests and the t-test
χ² test for independence
A test of H0\mathrm{H}_0: two categorical variables are independent, using a contingency table; it uses the upper tail only and needs every expected frequency greater than 55.Taught in 4.11 Chi-squared tests and the t-test