Generalized Linear Regression
This is a long post on an Intermediate level topic of interest to people working in Big-Data.
These are my notes, as I struggled to understand the topic from the available references. Unfortunately, the references, all contained the same wordings and the same areas of focus. They were deficient in some crucial areas – missing links
- To explain the Link function as the “Maximum Likelihood function” of the Original distribution
- To explain how Maximum Likelihood applied to regression – That the objective was to fit a probability distribution function, that maximized the probability that the observed independent variables would give as output the observed dependant variables
- That only a single pdf was being fit, even though the observations were in N-dimensions