Improving Detection Accuracy for Fake Customer Product Reviews
Open AccessOnline reviews are an important tool for guiding consumers’ eCommerce purchasing decisions. The eCommerce industry has seen explosive growth in online reviews for products and services have multiplied accordingly. Intense competition for the consumer dollar has led to an influx of fake reviews to lure or discourage potential buyers. By falsely influencing buying decisions, fake reviews result in dissatisfied customers, costly returns, and billions of dollars in wasted consumer spending. The ability to detect fake reviews with a high degree of accuracy is important to both online merchants and consumers but has posed technical challenges over the years. Current practices use machine learning (ML) models to analyze and classify review text. The detection accuracy of these methods is at best 87% and is sensitive to the inputs to the ML programs. In this research, natural language processing techniques were first used to preprocess the reviews and then ML models, including those using deep learning methods, were used to classify them. This research provides an improved method for increasing detection accuracy of fake reviews from 87% up to 89% compared to existing classification models, and our experiments demonstrate the efficacy and improved accuracy of the new model in identifying fake reviews.
- All rights reserved
Notice to Authors
If you are the author of this work and you have any questions about the information on this page, please use the Contact form to get in touch with us.