Bayesian classifier on top of Redis

Redis is a persistent, in-memory, key-value store with support for various data structures such as lists, sets, and ordered sets. All these data types can be manipulated with atomic operations to push/pop elements, add/remove elements, perform server-side union, intersection, difference between sets, and so forth.

Because of Redis' properties:

• It is extremely easy to implement simple algorithm such as bayesian filter.

• The persistence of Redis means that the Bayesian implementation can be used in real production environment.

• Even though I don't particularly care about performance at the moment, Redis benchmarks give me confidence that the implementation can scale to relatively large training data.

gem install bayes_on_redis

# Require BayesOnRedis and RubyGems
require "rubygems"
require "bayes_on_redis"

# Create instance of BayesOnRedis and pass your Redis information.
# Of course, use real sentences for much better accuracy.
# Unless if you want to train spam related things.
bor = BayesOnRedis.new(:redis_host => '127.0.0.1', :redis_port => 6379, :redis_db => 0)

# Teach it
bor.train "good", "sweet awesome kick-ass cool pretty smart"
bor.train "bad", "sucks lame boo death bankrupt loser sad"

# Then ask it to classify text.
bor.classify("awesome kick-ass ninja can still be lame.")

BayesOnRedis is also available in Python. With the same API.

easy_install bayes_on_redis

Fork http://github.com/didip/bayes_on_redis and send pull requests.