Page 162 - Kỷ yếu hội thảo khoa học lần thứ 12 - Công nghệ thông tin và Ứng dụng trong các lĩnh vực (CITA 2023)

146

Predicting Customer Sentiment with Vietnamese Hotel

Services by LSTM Model

2
1
1
1
Linh Bui Khanh , Anh Nguyen Quynh , Hai Tran Van , Ha Nguyen Thi Thu
1 Electric Power University, {linhbk, anhnq, haitv}@epu.edu.vn
2 Greenwich - FPT University, hantt194@fe.edu.vn

Abstract. The development of e-commerce has led to the strong growth of an
electronic hotel booking platform. Big data analytics in the hospitality industry
offers hotel managers and travelers many benefits. This study aims to predict the
sentiment of customers based on data from the TripAdvisor website to understand
customers' attitudes toward hotel services in Vietnam. Review data of customers
is collected from TripAdvisor with 22, 287 reviews of 12 hotels in 06 major cities
in Vietnam. An LSTM model is applied to train and predict customers'
sentiments. The results of this study with an accuracy of 96% show the
appropriateness of this model for predicting customer attitudes and satisfaction
toward hotel services in Vietnam.

Keywords: LSTM, customer review, Vietnamese hotel, data analysis, hotel
services, online review.

1 Introduction

The strong development of the Internet also leads to the development and growth of the
hotel industry [2, 4]. The hotel industry is a competitive market where hotel managers
are always looking for ways to provide the best quality in order to satisfy customers to
increase the return rate of guests [3]. Hotels are also adapting to different consumer
needs and developing different services and business models. Customer satisfaction in
the hotel industry today is a hot topic and essential in ensuring customer loyalty and
repurchase while building a good reputation and driving revenue for the hotel [5, 6].
Therefore, the research related to measuring customer sentiment in this field becomes
attractive to many scientists and hotel managers.
Under the influence of e-commerce and sharing economy, online booking sites are
also developing, recent studies are turning to data analysis of reviews that are generated
by customers on the Internet. The trend of studies from 2014 has diverted research
focuses on analyzing customer-generated big data on the Internet because of its high
availability, popularity, and large amount of data [7]. Data analytics offers the
opportunity to develop new techniques for mining and extracting meaningful value
from huge volumes of data [6]. The tourism and hospitality industries in many countries
strive to use data analytics to capture environmental changes and prepare long-term
plans and strategies. Unlike conventional approaches to a priori theories or hypotheses,
data analysis is a way of supporting data-driven decision-making [6]. One problem is

CITA 2023 ISBN: 978-604-80-8083-9