This sentiment analysis project aimed to predict the sentiment of tweets using the Sentiment140 dataset. The dataset was cleaned, and the text was transformed into numerical representations using the Bag-of-Words approach.
A Naive Bayes model was trained and evaluated, achieving accurate sentiment classification. The model's performance was enhanced through hyperparameter tuning. Users can now input their own tweets and obtain sentiment predictions as negative, positive, or neutral values.
In summary, this project successfully developed a sentiment analysis model for tweets with an overall accuracy of approximately 78.09% and precision for the positive sentiment of approximately 80.42%.