Ontology-based sentimental knowledge in predicting the product recommendations: A data science approach

Abstract

The effective item recommendations in the online business websites are largely based on the efficient study of product reviews. This paper describes the data science approach to predict product recommendations that are based on ontological sentiments calculated from consumer reviews. The recommendation strategy is based on the action taken on the discovered insight of deviated product recommendations obtained from the diagnosed machine learning model. It is revealed from the comparative results of Mean Absolute Error that there is vividness in the product recommendations in predicting the customer expected order of products when ontology support is used. This drives the customer purchase decisions on the product in a better manner.