Sentiment Analysis using logistic regression(from scratch).
The goal is to perform sentiment analysis of hotel reviews.
- Training data is in the form of positive and negative hotel reviews.
Features:
- There are 6 following features selected for this project:
- Consider following words for feature 4: "I", "me", "mine", "my", "you", "your", "yours", "we", "us", "ours".
Data:
- There are two files pos_reviews.txt and neg_reviews.tx which contains positive and negative reviews respectively.
- pos_words.txt and neg_words.txt files contain positive and negative words respectively.
Code:
- There are two files in code: feature_extractor.py and logistic_regression.py.
- feature_extractor.py extracts the above mentioned features from the Hotel reviews available and logistic_regression.py performs Logistic Regression on those features. It is implemented from scratch without using any scikit-learn libraries.
Result:
Conclusion: