Employing Deep Learning Methods for Predicting Helpful Reviews
Abdalraheem Alsmadi, Shadi Mahmoud Faleh AlZu’bi, Bilal Hawashin, Mahmoud Al‐Ayyoub, Yaser Jararweh · 2020
E-commerce dominates a large part of the world's economy with many websites dedicated to selling products online. The vast majority of e-commerce websites provide their customers with the ability to express their opinions about the products/services they purchase. These reviews represent a rich source of information about the users' experiences, which is of great benefit to both the producer and the consumer. In This paper we present a set of machine/deep learning models, especially using Recurrent Convolutional Neural Network (RCNN) to predict the helpfulness of reviews. Mainly, two approaches are used: a supervised learning approach and a semi-supervised approach. The latter is a unique aspect of our work and it takes advantage of a large number of unlabeled reviews. The results show that both approaches are better than existing approaches. Moreover, the results show that the second approach has a remarkably better performance compared with the first one, which is in accordance with recent trends in machine/deep learning that focus on benefiting from the huge amount of unlabeled data to enhance the performance of supervised models.