Designed a powerful algorithm, which pulls tweets by topics and has a custom NLP trained model for sentence modeling using many methods such as “Bag of Words” method, to accurately predict the sentiment of a tweet and confirm the entity’s in the sentence. This algorithm can be used in moderating tweets, using tweets along with financial data to predict impact on stocks, finding people in need of help, and just about anything that you can think that exists on social media.
This algorithm is just a proof-of-concept and using NLP model based on only a single parameter, in production, many other parameters such as frequency of tweets, sentiment of tweets, or maybe a collection of tweets in the last couple of days, can be included to boost the robustness of the model, and provide new insights.
python sentiment_analysis.py