Building a Social Media Sentiment Analysis Model | | FULL Tutorial

The video shows how to build sentiment analysis model using logistic regression. The sentiment140 dataset is used for training the model. When any sentence is fed to the model, it predicts weather the sentiment is positive or negative. I used regex & nltk for tweets cleaning and tf-idf vectorizer (from sklearn) for vectorizing and logistic regression to train the model. Few mistakes I'd like to correct from the video: 1. The weights or input vectors cannot be called as "tensors". 2. In sigmoid function graph, I said "tends to infinity", its acutally "tends to 1" I had to trim the part of making a classification report, model comparisons, bar graph, wordcloud, etc. as the video got too lengthy. The github link for this project :