Tweet Classifier

Java-based Tweet classifier based on Mallet's Naive Bayes classifier. Currently restricted to a limited number of categories:

• Personal Business
• Work / School
• Social / Recreation
• Other

InferClass

####Parameters:####

• t = The text content of a tweet. Must be greater than 50 characters.

####Output:####

A JSON Object containing the classification values based on the classes indicated above.

####Example:####

http://~/InferClass?t=I am so tired of my job, I can't wait to work at a new place where everyone is friendly

UserInferClass

####Parameters:####

• username = Twitter username
• page (optional) = Twitter feed page of content
• count (optional) = Twitter feed number of tweets per page. Keep in mind that the actual number may be smaller if there are tweets with less than 50 characters (will not be classified)

####Output:####

A JSON Object containing the username and an array of tweet objects. The tweet objects contain a date/time value as well as an array of classification values based on the classes indicated above.

####Example:####

http://~/UserInferClass?username=grantdmckenzie&page=2&count=13

TODO:

• Better Classification based on more than single user manual classification.
• Add "Shopping" as category.