Due to the digitalization, there exists an increased amount of user-generated content on the Internet, where people express their opinions on various topics. Sentiment analysis is the statistical and analytical examination of human emotions and opinions regarding a certain subject. Our study extends the literature by developing a data analytics methodology for the benchmarking of sentiment scoring algorithms in the context of online customer reviews. We demonstrate the applicability of the methodology using Amazon product reviews as the source data. Analyzing text-based content such as Amazon customers’ reviews through text analytics and sentiment analysis can help Amazon and other online retailers to discover valuable actionable insights regarding their products. The contributions of this study are twofolds: to examine the predictive power of machine learning (ML) algorithms with respect to predicting sentiment scores and to analyze patterns in the differences between scores obtained from different sentiment scoring algorithms.

Please click here to download the final draft of the article.

Please Cite as Follows:

Abou-Kassem, T., Alazeezi, F.H.O., Ertek, G. (2023). A Data Analytics Methodology for Benchmarking of Sentiment Scoring Algorithms in the Analysis of Customer Reviews. In: Yang, XS., Sherratt, R.S., Dey, N., Joshi, A. (eds) Proceedings of Eighth International Congress on Information and Communication Technology. ICICT 2023. Lecture Notes in Networks and Systems, vol 693. Springer, Singapore. https://doi.org/10.1007/978-981-99-3243-6_46