Statistics-Formula Sheet Final Exam Uniformin the interval [a, b]: 1 a<x<b f (x) = b − a , 0 otherwise 0 x≤a x − a a<x<b F (x) = b−a 1 x≥b Basics Counting Permutations: n! . n Pk = (n − k)! E[X] = Counting Combinations: n! n . n Ck = k = k!(n − k)! a+b 2 , V (X) = (a−b)2 12 Exponential pdf: f (x; λ) = λe−λx , E[X] = Gamma pdf: f (x; α, β) = V (X) = αβ 2 . Conditional Probability Formula: P [A ∩ B] P [A|B] = . P [B] Bayes Theorem (with partition B1 , · · · , Bk ): P [Bi ] × P [A|Bi ] P [Bi |A] = k . X P [Bj ] × P [A|Bj ] 1 1 , V (X) = 2 . λ λ 1 α−1 e−x/β , β α Γ(α) x E[X] = αβ, Normal pdf: 1 x−µ 2 f (x) = σ√12π e− 2 ( σ ) , −∞ < x < ∞, E[X] = µ, V (X) = σ 2 . j=1 Chi-square pdf: 1 f (x, ν) = 2ν/2 Γ(ν/2) xν/2−1 e−x/2 , E[X] = ν, V (X) = 2ν. Discrete Uniform: Two Variables Z ∞ p(x; N ) = 1/N for x = 1, · · · , N . E[X] = 2 Marginal Density of X: fX (x) = f (x, y)dy (N + 1)/2, V (X) = (N − 1)/12. Random Variables −∞ Binomial: b(x; n, p) = np, V (X) = npq. Poisson: p(x; λ) = λx −λ , x! e n x px (1−p)n−x , E[X] = Conditional Density of X given Y = y: f (x, y) fX|Y (x|y) = fY (y) Covariance: Cov(X, Y ) = E[XY ] − E[X]E[Y ] E[X] = λ, V (X) = λ. Hypergeometric: h(x; n, M, N ) = V (X) = n · M x N −M n−x N n , E[X] = n · Correlation: M , N Corr(X, Y ) = ρ(X, Y ) = M M N −n · (1 − )· . N N N −1 Cov(X, Y ) σX σY Confidence Intervals and Testing: Population Mean:√ √ Neg. binom pmf: C.I: x̄ ± (zα/2 )σ/ n Or x̄ ± (tα/2 )s/ n r(1 − p) r x N B(x; r, p) = x+r−1 , X̄ − µ0 r−1 p (1−p) , E[X] = p √ Test Statistic: Z = σ/ n r(1 − p) V (X) = . p2 X̄ − µ0 √ , with n − 1 d.f. T = Geometric pmf: s/ n 1 x−1 g(x; p) = p(1 − p) , E[X] = , V (X) = p Sample Size Calculation: (1 − p) When α, σ and the width w of the C.I. are all . p2 known: 1 n= 2z α2 σ 2 w . Population Proportion: r p̂q̂ C.I.: p̂ ± zα/2 , n Test Statistic: p̂ − p0 Z=r . p 0 q0 n Sample Size Calculation: For C.I.: When α and the width w of the C.I. are all known: !2 2z α2 p(1 − p) . n= w Sample Size Calculation in Hypothesis Testing: (When both α and β(µ0 ) are given) σ(zα + zβ ) 2 n= µ0 − µ0 OR σ(zα/2 + zβ ) 2 n= . µ 0 − µ0 2