A Chi-squared distribution with statistical degrees of freedom is a Gamma distribution with shape and rate . Its Probability density function isFor a positive integer , a variable is the sum of squares of independent standard normal random variables. The gamma definition extends the law to noninteger positive degrees of freedom; it does not require a fractional number of Gaussian coordinates.
Its mean is , its variance is , and its Moment-generating function is for . These follow directly from the Gamma density or by multiplying the squared-normal transforms in the integer case.
The law appears when Gaussian deviations are aggregated as a squared length. For iid observations with and , the Sample variance satisfies . Centering removes the one-dimensional sample-mean direction, leaving independent Gaussian coordinates after an orthogonal change of basis. This exact normal-sampling law underlies confidence intervals for a variance; it is not a universal variance-sampling formula for arbitrary data.
For independent centered normal random variables of common variance , has a chi-squared distribution with degrees of freedom. Its density gives the displayed moments by a gamma integral. In three dimensions, . Radial odd moments need not vanish, although any odd Cartesian product does.
A Chi-bar-squared distribution is a mixture distribution , where and , and denotes a point mass at zero. It arises when a Gaussian likelihood is optimized under inequality constraints: projection onto a constraint cone selects different dimensions according to which inequalities are active. For example, if are independent standard Gaussian random variables, then has distribution . On it is ; on it is . Symmetry makes the sign of independent of , proving the stated weights and component laws.
If are independent standard normal random variables and , then has a Noncentral chi-squared distribution with degrees of freedom and noncentrality . For it reduces to the chi-squared distribution. Squaring a shifted signed normal Wald statistic gives the case .
For with , completing the square in its Gaussian representation gives . Integrate the identity and use Tonelli theorem to obtain the displayed reciprocal moment. It is strictly decreasing in , since its positive integrand is strictly decreasing for every . At the integral equals , proving finiteness and the sharp maximum over noncentrality parameters.
For and , the upper tail is at most above , and the lower tail is at most below . Indeed for . The Chernoff bound with proves the upper estimate. For the lower estimate, for , and optimization gives the result. The lower threshold can be negative; the chi-squared Chernoff lower-tail bound provides an always-positive alternative.
For and , apply the Chernoff bound to using . Optimization at proves the displayed estimate. It remains useful when the additive lower threshold in the chi-squared concentration inequality is nonpositive. For , choosing gives a tail at most because .
A sum of the squares of independent random variables with the standard normal distribution has the chi-squared distribution with degrees of freedom. A scaled Gaussian integral and factorization by independence give its moment-generating function for . For positive , this expectation diverges when .
For parameters and , the Scaled inverse chi-squared distribution is the law ofThe inverse change of variables gives its normalized densityIt is an Inverse-gamma distribution with shape and scale . The label is a scale parameter, not the variance of .
Its mean is for , and its variance is for . Below those respective thresholds the positive moments diverge. For example, gives mean three and variance nine.
This is a conjugate prior for the unknown variance of a Normal distribution with known mean . With prior parameters and independent observations, the posterior has and . Multiplying prior and likelihood adds their powers and inverse-variance exponents, which proves the update. An unknown estimated mean requires a separate model and changes the update.
Chi-squared distributions are stochastically increasing in their statistical degrees of freedom. For any real , take independent and . Gamma additivity givesFor integer , this is the familiar construction by adding extra independent squared standard normal variables. For noninteger degrees of freedom, multiplying the gamma moment-generating functions proves the same statement.
Under this coupling the larger-degree variable is larger on every sample path. Hence for every , and in fact the inequality is strict for : average and use and the strictly positive chi-squared density. Therefore every quantile at a fixed probability in strictly increases with the degrees of freedom. Comparing chi-squared critical values must account for this change; the distribution is not fixed while only its label changes.
A Chi-squared asymptotic approximation uses a Chi-squared distribution as the limiting null law of a statistic. In a regular identifiable fixed-dimensional model, Wilks theorem gives for smooth independent null restrictions, provided the true parameter is an interior point and the usual differentiability and information conditions hold. This is a limit as Sample size grows, not an exact finite-sample identity.
For a fully specified multinomial model with a fixed number of categories and probabilities , Pearson's statistic isThe constraint removes one Gaussian degree of freedom. Its usual fixed-category assumptions make all expected counts grow. Growing dimension or persistent rare counts require separate approximation analysis; merely choosing chi-squared critical values does not supply it.
Articles by others on the same topic
There are currently no matching articles.