Parameter Estimation And Hypothesis Testing Of The New Panel Model

The Panel data model is one of the hot spot model in the statistical analysis of models, at present. Existing Panel model which is researched mainly is about of one or two main effects. In order to study more complicated data accurately, this paper established a new Panel data model which has three main effects and interaction effects, a total of eight random effects. First of all, this paper gives a new Panel model expression and the expression of covariance matrix by the spectral decomposition method. Secondly, using the least square unified theory, the paper give the fixed effects parameter estimation which is least squares estimation (LSE), generalized least squares estimation (GLSE), two-stage estimation and the simple proof of the unbiasedness. In terms of variance component, the paper use generalized p-value theory to solve hypothesis test problems that has redundant parameters and established exact tests and confidence intervals.