© 2015-2016 Matthew Cunningham

This is a simple naive bayesian classifier to gain independent probabilistic assumptions on test input. The classifier requires precisely 2 groups with training data. This is just a fun side project I did over the weekend, but any contributions would be fantastic.

There are three files in this repository — each file works on its own:

• bayes.min.apl - This is the completely minified version.
• bayes.apl - This file isn't as obfuscated as the minified version, but it still follows conventional minified APL practices.
• bayes.full.apl - Fully documented

This example shows the classifier displaying the independent probabilistic assumptions on whether given text aligns with keywords of two different animals: cat or dog.

$ ("meow purr hiss bad animal" 0) ("bark grr howl good animal" 1) b "I am an animal and I hiss"
0.9999999995 4.999999994999999e¯10

Try out this example using ngn/apl's online interpreter.

The provided training data is the keywords and groups, with cat being the 0th index and dog being the 1st index. The test data consists of a string that contains keywords to align with a certain training group; in this example, cat is the classified group.