Sentiment Analysis of Social Media Reviews using Machine Learning and Word Embedding Techniques

The enormous amount of data generated and exchanged daily on social media raises the need for knowledge-extraction tasks to be mechanized. Sentiment analysis is an active area of information extraction research with various challenges to overcome. Any business ought to be highly interested in client feedback. Customers mainly rely on reviews to determine where to dine. Sentiment analysis is vital in classifying restaurant reviews into positive, negative, and neutral polarity to determine whether the cuisine is good, securely produced, and worth picking over other restaurants. The Yelp dataset for restaurant reviews is used in this study to test different word embedding approaches, including Bag of Words (BoW), Term Frequency-Inverse Document Frequency(TF-IDF), GloVe, Word2Vec, and Doc2Vec. Supervised Machine Learning (ML) algorithms like Logistic Regression (LR) and Support Vector Machine (SVM) are evaluated based on the performance metrics such as the F1-Score, Accuracy, Recall, and Precision. Comparative findings demonstrate that combining a Support Vector Machine and TF-IDF word embedding technique produced results with an accuracy of 98%.