enzoseason / mult-class-text-classification

Mult-class text classification

Project of classifying tweets of company Limportant into 13 categories

1. Highlights:

• Language: Python
• It uses CountVectorizer, TfidfTransformer to pre-processing text data
• Machine learing package is sklearn
• It uses GridSearch to get hyperparameter of the best model
• Ensemble modeling with voting
• Full-score, which combines precision and recall, is 0.68/1.00

2. Data:

2.1 Training:

• Training size: 18 000 tweets
• Input: text, category_id
• Output: category_id

2.2 Predicting:

• Input: text
• Output: category_id

3. Reference:

• Multi-Class Text Classification with Scikit-Learn