Abstract— In this paper, ML strategy for detecting phoney reviews is proposed. Using a range of feature engineering methodologies, this paper extends beyond the review's feature extraction technique to identify various examiner behaviours. The paper compares and contrasts the outcomes of various programmes. We compared the effectiveness of several classifiers in both cases using an actual dataset of restaurant ratings with and without factors derived from the user's activitiesWhen it comes to the f-score, the remaining classifiers have the highest f-score of 74.79 percent. According to the results. The f-score has grown by 14.14 percent after factoring behavioural characteristics of the reviewers.