This NLP Insights Analyzer project leverages a combination of text preprocessing techniques, including stemming, stop-word removal, and CountVectorizer, to transform raw customer reviews into a format suitable for machine learning models. The project explores various classification models, including Logistic Regression, Naive Bayes, Support Vector Machines, and Random Forests, to predict and categorize sentiments.
The models are evaluated using cross-validation to ensure robust performance across different datasets. The best-performing model is then saved and can be easily loaded for making predictions on new, unseen data.
This project used the "ClassifyReviews_NLP/Restaurant_Reviews.tsv" sourced from Github. The dataset contains a collection of reviews labeled as positive and negative for training and testing the classifier.
Link: https://github.com/PritiG1/ClassifyReviews_NLP/blob/main/Restaurant_Reviews.tsv