Maximum Likelihood Theory for Log-Linear Models
This chapter presents the basic theoretical results of fitting log-linear models by maximum likelihood. The level of mathematical sophistication is considerably higher than in the rest of the book. The presentation assumes knowledge of advanced calculus, mathematical statistics, and large sample theory. Although the results in this chapter are proven in a different manner than for regular linear models, the results themselves are quite similar in nature. The common linear structure of the two techniques leads to the well-known analogies between them. A familiarity with log-linear models at the level of, say Fienberg (1980), is assumed.
KeywordsLogit Model Asymptotic Property Asymptotic Distribution Saturated Model Moment Generate Function
Unable to display preview. Download preview PDF.