Definitions
Probability distributionsDiscrete: probability distribution, cumulative probability distribution. Continuous: cumulative distribution function, probability density function (p(x) = limit of Prob(x < X < x + Δx)∕Δx as Δx → 0). Moments: Means: ∑ xi∕N = ∑ count(x)x∕N = ∑ p(x)x (discrete), ∫ p(x)xdx. Variances: ∑ p(x)(x -)2, ∫ p(x)(x -)2dx. Higher moments: skew, kurtosis. Other descriptors: median, mode.
BestiaryPretty good summaries on Wikipedia (http://en.wikipedia.org/wiki/List_of_probability_distributions). R help pages. Books: [1, 2], Johnson, Kotz, Balakrishnan et al. Characteristics (discrete vs continuous; range (positive, bounded, …); symmetric or skewed …)
(Insanely thorough version: [3].)
Jensen’s inequalityBottom line: E[f(x)]≠f(E[x]), unless f is linear (need new notation for expectation over a particularly probability density function p(x): Ep[f(x)] ≡∫ p(x)f(x)dx. Quantify exactly by calculating the expectation precisely, or by the delta method. Delta method: Ep[f(x)] ≈ Ep[f()] + Ep[f′(x)|x=(x-)] + 1∕2Ep[f′′(x)|x=(x-)2] = f() + 1∕2f′′()Var(x).
References
[1] B. M. Bolker. Ecological Models and Data in R. Princeton University Press, 2008. [2] M. Evans, N. Hastings, C. Forbes, and J. B. Peacock. Statistical Distributions. John Wiley & Sons, 2010. [3] L. M. Leemis and J. T. McQueston. Univariate distribution relationships. The American Statistician, 62(1):45–53, 2008. |