hiteshwagle/Amazon_Product_Review_Similarity

Performed similarity analysis using PySpark based on given positive and negative reviews. Reviews are first divided in Negative(<3 rating) and Positive(>3 rating) categories based on given ratings. Google Pre-trained universal sentence encoder and Tf-Idf in Spark algorithms are implemented to find Intra class similarity for Positive and Negative reviews and to find class center sentence for each positive and negative category.