Analysing and predicting the online review helpfulness: Using machine learning and meta-heuristic algorithms (The case study of TripAdvisor website)

As relationship marketing promotes products through word-of-mouth advertising, online consumer reviews can work as an alternative in developing modern digital businesses. In this research, four online review helpfulness measurements were employed for online review helpfulness prediction, using machine learning and meta-heuristic algorithms. The results indicated the most powerful algorithm for online review helpfulness prediction was the light gradient boosting machine regression. Also, the decimal logarithm of the number of helpful votes is the most suitable scale among online review helpfulness measurements. At last, the top 20 predictors were extracted from the two most accurate algorithms and the most efficient measurements. 2024 Taylor & Francis Group, LLC.