Skip to content

HTTPS clone URL

Subversion checkout URL

You can clone with
or
.
Download ZIP
Bayesian classifier on top of Redis
Python Ruby
tree: a54c4ea1ee

Fetching latest commit…

Cannot retrieve the latest commit at this time

Failed to load latest commit information.
config
python
ruby
.gitignore
README.markdown

README.markdown

What is BayesOnRedis?

Bayesian classifier on top of Redis

Why on Redis?

Redis is a persistent in-memory database with supports for various data structures such as lists, sets, and ordered sets. All this 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.

How to install?

gem install bayes_on_redis

Getting started

# 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 => 5)

# 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.")

for Pythonistas

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

Contributing

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

Something went wrong with that request. Please try again.