In this project I have worked on class-imbalance problem of machine learning. Class-imbalance problem occurs when there is a difference between majority and minority class. For selecting the samples from the majority class I have used meta heuristic optimization algorithm named Particle Swarm Optimization (PSO) and Ring Theory based Evolutionary Algorithm (RTEA).
In this project a news recommendation system is built with the help of MIND dataset which is dynamic in nature and varies from person to person. Here Tf-idf, LSTM and BERT models are used.
Here I have used both supervised and unsupervised learning techniques to get the sentiment of the text. Supervised learning methods are rule-based model VADER and another algorithmic NLP based sentiment analysis model. and logistic regression is used in unsupervised learning model. Then the users are seperated in various groups according to their review and same type of users are clustered in same group.