sentimentClassifier

A vanilla Naive Bayes classifier trained with tweets that tells you whether or not you are saying something nice.

Like this:

Tell me something...
love
That is nice.

Tell me something...
hate
That is not nice.

Tell me something...
homework
That is nice.

Tell me something...
homework on friday
That is not nice.

See the code here.

Run it by either copying the source code and building it with matrix and random as dependencies; or by cloning this repository:

0. Install stack if necessary.

1. Clone repository:

   git clone https://github.com/da-cali/sentimentClassifier
   

2. Open folder:

   cd sentimentClassifier
   

3. Build project:

   stack build
   

4. Train and run:

   stack runhaskell src/Main.hs
   

Note: Increase the sample size to train with more examples and get more accurate responses.

Authors:

Louise Brett, Dan Castillo, Michael Ton.