The purpose of this work is to create an automated text-based categorization system that can predict the usefulness of Amazon online consumer reviews with high accuracy. The challenge at hand is to do a binary classification utilising a combination of text-based features and machine learning algorithms. The binary classifications will be specified as follows: ‘1’ denotes ‘Helpful’ and ‘0’ denotes ‘Not helpful.’ The number of users who voted the review as ‘helpful’ will be used to compute the usefulness score. This study will employ a variety of text-based features, including matrix-based vectorized features, word embedding-based features, and structural, syntactic, semantic, and metadata features derived from review and summary tex