Modelling customer satisfaction from online reviews using ensemble neural network and effect-based Kano model

Abstract

With the rapid advances in information technology, an increasing number of online reviews are posted daily on the Internet. Such reviews can serve as a promising data source to understand customer satisfaction. To this end, in this paper, we proposed a method for modelling customer satisfaction from online reviews. In the method, customer satisfaction dimensions (CSDs) are first extracted from online reviews based on latent dirichlet allocation (LDA). The sentiment orientations of the extracted CSDs are identified using a support vector machine (SVM). Then, considering the existence of complex relationships among different CSDs and the customer satisfaction, an ensemble neural network based model (ENNM) is proposed to measure the effects of customer sentiments toward different CSDs on customer satisfaction. On this basis, to identify the category of each CSD from the customer’s perspective, an effect-based Kano model (EKM) is proposed. Finally, an empirical study, which consists of two parts (phones and cameras), is given to illustrate the effectiveness of the proposed method.

Keywords:

• customer satisfaction
• online reviews
• ensemble neural network
• Kano model
• product improvement

ORCID

Jian-Wu Bi http://orcid.org/0000-0003-2253-3492

Yang Liu http://orcid.org/0000-0002-5113-8638

Zhi-Ping Fan http://orcid.org/0000-0001-6778-4637

Additional information

Funding

This work was partly supported by the National Natural Science Foundation of China [project numbers 71771043 and 71871049], Liaoning BaiQianWan Talents Program [project number 2016921027], the Fundamental Research Funds for the Central Universities, China [project number N170605001], and the 111 project [B16009].