abdulfatir/twitter-sentiment-analysis

Implements a comprehensive comparative framework supporting 10+ classification algorithms (Naive Bayes, SVM, Decision Trees, Random Forest, XGBoost, Logistic Regression, MLP, LSTM, CNN) with standardized CSV input/output format and optional bigram feature extraction. Includes preprocessing pipelines using NLTK for text normalization and frequency distribution analysis, with Keras/TensorFlow backends for deep learning models and scikit-learn for traditional classifiers. Supports ensemble methods through majority voting and CNN feature extraction for downstream SVM classification.

1,643 stars. No commits in the last 6 months.