The Internet accompanies us in every moment of our lives, supporting us in many ways. Among these, it helps us when we want to choose the best services and products. So when it comes to picking a movie to watch, a restaurant to eat in, a hotel to stay in, or a product to buy, we grab our smartphone and visit one of the countless sites where users can report their experiences and read those of others. However, as often happens, even in this case there are possible scams of which we must beware. In this paper, we propose a machine learning system for predicting the reliability of online reviews. Specifically, our system collects reviews from Amazon, extracts various features, and gives them as input to an ensemble learning system based on three anomaly detection algorithms. To demonstrate the benefits of our approach, we report the results of a comparative analysis with some state-of-the-art systems using the data collected by ReviewMeta. These results have allowed us to realize how widespread the phenomenon of online fake reviews is.
Sansonetti, G., Gasparetti, F., Micarelli, A. (2023). A Machine Learning Approach to Prediction of Online Reviews Reliability. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) (pp.131-145). Cham : Springer Science and Business Media Deutschland GmbH [10.1007/978-3-031-35915-6_11].
A Machine Learning Approach to Prediction of Online Reviews Reliability
Sansonetti G.
;Gasparetti F.;Micarelli A.
2023-01-01
Abstract
The Internet accompanies us in every moment of our lives, supporting us in many ways. Among these, it helps us when we want to choose the best services and products. So when it comes to picking a movie to watch, a restaurant to eat in, a hotel to stay in, or a product to buy, we grab our smartphone and visit one of the countless sites where users can report their experiences and read those of others. However, as often happens, even in this case there are possible scams of which we must beware. In this paper, we propose a machine learning system for predicting the reliability of online reviews. Specifically, our system collects reviews from Amazon, extracts various features, and gives them as input to an ensemble learning system based on three anomaly detection algorithms. To demonstrate the benefits of our approach, we report the results of a comparative analysis with some state-of-the-art systems using the data collected by ReviewMeta. These results have allowed us to realize how widespread the phenomenon of online fake reviews is.I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
