Amazon products reviews classification based on machine learning, deep learning methods and BERT

In recent times, the trend of online shopping through e-commerce stores and websites has grown to a huge extent. Whenever product is purchased on an platform, people leave their reviews about product. These are very helpful for store owners product’s manufacturers betterment work process as well quality. An automated system proposed in this that operates two datasets D1 D2 obtained from Amazon. After certain preprocessing steps, N-gram word embedding-based features extracted using term frequency-inverse document frequency (TF-IDF), bag words (BoW) global vectors (GloVe), Word2vec, respectively. Four machine learning (ML) models support vector machines (SVM), logistic regression (RF), (LR), multinomial Naïve Bayes (MNB), deep (DL) convolutional neural network (CNN), long-short memory (LSTM), standalone bidirectional encoder representations (BERT) used classify either positive or negative. The results by standard ML, DL BERT evaluated performance evaluation measures. turns out be best-performing model case with accuracy 90% derived embedding while CNN provides best 97% upon D2. shows better overall compared D1.