Recommendation System Based on Sentiment Analysis

Built an end-to-end sentiment analysis pipeline leveraging

NLP preprocessing

(tokenization, stop-word removal, stemming) on

scraped

Twitter and Google Maps reviews.

•

Engineered features and evaluated multiple ML models (

SVM, Decision Tree, Logistic Regression, Naive Bayes

) to classify sentiment into

positive, negative, and neutral

categories.

•

Optimized model performance through hyperparameter tuning and comparative evaluation

•

Demonstrated expertise in

data collection, text preprocessing, model experimentation, and evaluation

.