|
Empirical likelihood is Owen (1988) proposed a non-parametric inference method , followed by some scholars to apply it linear model , semi- parametric model , regression function , density kernel estimator , biased samples , etc., but these are mostly independent samples discussed under the same distribution . CUI Heng Jian ( literature [ 1 ] ) Suppose X 1, LXn as iid samples taken from a population of X and θ0 = EH (X, μ), where μ is nasty argument , discussed the experience interesting parameter θ0 likelihood Ratio confidence Intervals article will condition θ0 = EH (X, μ) extended to the general estimating equations EH (X, θ0, μ) = 0 and the sample {X i | i ≥ 1} is strongly stationary φ mixing sequence ( defined see ref [ 2 ] ) the situation , using the empirical likelihood ratio method to construct a confidence region θ0 Let H (x, s, t) is a continuous multivariate function and note Θμδ = {t | t ∈ R, | t-μ | ≤ δ}, where δ gt; 0. the general estimation equation EH (X, θ0, μ) = 0, where X is a random variable , θ0, μ is the unknown parameters , to obtain θ0 confidence region , first take μ Consistency estimate μ? (X 1, L Xn) =? μ?, so its empirical likelihood ratio =??? Π ≥ Σ = Σ =???
|