A Novel K-Means Clustered Support Vector Machine Technique for Prediction of Consumer Decision-Making Behaviour

Abstract

A greater number of consumers are using social networks to express their feedback about the level of service provided by hotels. Online reviews from patrons can be used as a forum to enhance the level of service of hotels. Customer reviews are indeed a reliable and dependable source that aid diners in determining the quality of their cuisine. It is critical to develop techniques for evaluating client feedback on hotel services. In order to accurately anticipate the consumers' decision-making behaviors based on hotel internet evaluations, this study proposes a novel K-Means Clustered Support Vector Machine (KMC+SVM) technique. Principal Component Analysis (PCA) is employed to determine the characteristics from the preprocessed data while the Min-Max normalization approach is used to standardize the raw data. The performance of the suggested technique is then evaluated and contrasted with a few other methods that are currently in use in terms of accuracy, sensitivity, RMSE, and MAE. The findings demonstrated that segmenting customers based on their online evaluations can accurately predict their choices and assist hotel management in establishing priorities for service quality enhancements.