Data Mining Models in SQL
Gordon S. Linoff · 2015
Data mining is the process of finding meaningful patterns in large quantities of data. SQL provides a good basis for learning the basics about data mining. This chapter introduces important data mining concepts and the processes of building and using models. It takes an alternative approach that introduces data mining concepts using databases. The first type of model in this chapter is the look-alike model, which takes an example, typically of something particularly good or bad-and finds other rows similar to the example. The next type of model is the lookup model, which summarizes data along various dimensions to create a lookup table. Lookup models lead to naïve Bayesian models, a powerful technique that combines information along any number of dimensions, using some interesting ideas from the area of probability. Directed data mining, tasks that models might be used for, and methods of evaluating models are also discussed.