Online Fake Review Detection based on Multiple Feature using Machine Learning Technique

In today's online era while purchasing things such as clothes, laptops, home appliances, etc., customers rely on other people’s reviews to know more about the product which they ought to buy. These reviews can be genuinely posted by previous purchasers or be totally deceptive i.e. spammed by bots. Therefore, the detection of fake reviews online has now become the pinnacle point of interest. Despite recent studies, the accuracy of the existing models is not up to the mark. The proposed method discusses the online fake review de- tection consisting of data acquisition, pre-processing, train-test splitting of data, and ensemble learning which detects these fake reviews and increases the accu- racy to a higher mark. Our model uses three supervised machine learning algo- rithms viz. Support Vector Machine (SVM), Random Forest (RF), Artificial Neural Network (ANN), and trains on electronic items dataset. After training on three individual algorithms and post-calculation of the accuracy of each algo- rithm, the model compares and analyzes the outcomes and selects the algorithm with the highest performance accuracy.