Applying natural language processing to analyze customer satisfaction

Abstract

The aim of this paper is to analyze customer satisfaction by applying natural language processing (NLP). We have collected over 50,000 airline reviews from TripAdvisor data in the period from 2016 until 2019. This analysis demonstrates the capability of discovering the pain points of the customers by using data science techniques related to NLP. Our study shows that in today`s world, data-driven decisions must be taken quickly in order to maintain customer satisfaction and prevent customer churn.

Evaluation Model and Influencing Factors of Consumer Satisfaction with E-Commerce Platform

“…In the EIS of consumer satisfaction with e-commerce platform, the correlation intensities between overall satisfaction and primary indices are denoted as S (1) i(i=1, 2, …, n); the importance of primary indices is denoted as φ (1) i (i=1, 2, …, n); the autocorrelation intensities between primary indices is denoted as sjn(i=1, 2, …, n); the correlation intensity between primary index i and secondary index j is denoted as Sij(i =1, 2, …, n; j =1, 2, …, m); the importance of secondary indices is denoted as φ (2) j(j=1, 2, …, m); the satisfaction indices of primary indices are denoted as SA (1) i(i=1, 2, …, n); the consumer satisfaction evaluated by secondary indices is denoted as ERj(j=1, 2, …, m).…”

“…The evaluation model for consumer satisfaction with ecommerce platform can be fuzzy normalized by formulas ( 26) and ( 27). After fuzzy normalization of S (1) i and sin (i=1, 2, n), the upper bound (S'ij) V α and lower bound (S * ij) K α of the α-cut set of S'ij * can be calculated. If α=1, there exists:…”

“…In the evaluation model for consumer satisfaction with ecommerce platform, the importance of secondary indices is jointly determined by the fuzzy normalized relationship between primary and secondary indices, and the importance of primary indices φ * (1) i(i =1, 2, …, n). Let S'ij * be the fuzzy normalized relationship between primary index i and secondary index j.…”

“…Next, the mean fuzzy triangular number ER * j is crossmultiplied with the fuzzy coefficient Q *(2) j of secondary index importance at different levels of α to obtain the consumer satisfaction SA * (1) i (i=1, 2, …, n) with primary indices at different levels of α:…”

“…With the rapid development of e-commerce related technologies, shopping on e-commerce platform has become a main consumption mode [1,2]. The intense competition between e-commerce platforms motivates e-commerce websites to enhance their service quality by improving shopping experience and consumer satisfaction [3,4].…”