Machine learning and Redis

While the hype cycle continues for what is generally called "Big Data", Redis offers numerous ways to actually accomplish some of what the advertising and media is promising to business users and leaders. Besides being a good choice for performing quick-and-dirty loading and manipulation of data, Redis also performs well as a staging platform for data in a transitional mode that is later manipulated towards a final state, depending on the application. Redis use as a datastore in machine learning techniques and approaches helps as an easily malleable store supporting a particular learning algorithm.

This section takes two supervised learning tasks, Näive Bayes and linear regression, to demonstrate different approaches to statistical analysis with Redis as a transitory datastore for intermediate results. For the first example, the dataset is a pre-existing set of 52 MARC21 records for Jane Austen's Pride and Prejudice and Herman Melville's Moby Dick. This dataset...