Aim Sentiment Analysis of Tweets in Hinglish language and classify them as normal, abusive or hateful by applying Naive Bayes Classifier.
Data set has been Collected from twitter using tweepy twitter api. The collected datasets has been manually labelled and then we performed following things:
1. Pre-processing of Datasets: The tweets were first processed and the unnecessary information were deleted from the tweets like @ and symbols and others characters of length 3. Hash Tag were converted into normal words. Link were removed from tweets
2. Tokenized: The tweets were tokenized to split them word by word. 3. Model Building: From the processed tweets we created a Machine Learning Model for Naive Bayes Classifier. We split the data datasets into training and Testing Sets.(75-25 %) 4. Prediction And Result: We then test our model and perform the prediction on the testing sets and calculated the results. We got an accuracy of 76.35327635327636.