ViugiNick / SocialMediaLab1

SocialMediaLab1

Console application for tweet classification:

Modes

1. sentiment - Sentiment classification using tweets
2. company - Company classification using tweets
3. mixed - Sentiment classification using tweets and companies

Usage

You can specify train and test files (--train, --test options) and mode(--mode option)

After running the application, just select classifier and follow the instructions

type 'exit' to stop application

Usage example

!python3 main.py --mode=sentiment
Select classifier:
1 - SentiWordNet dictionary classifier
2 - RandomForest classifier
3 - Bayes classifier
3
You selected Bayes classifier
Classifier accuracy:
0.7510548523206751
             precision    recall  f1-score   support

negative       0.75      0.49      0.59        49
neutral        0.77      0.93      0.84       156
positive       0.56      0.28      0.38        32

avg / total    0.74      0.75      0.73       237

[[ 24  22   3]
 [  7 145   4]
 [  1  22   9]]

Enter your tweet:
What a nice example
positive