Page 167 - Kỷ yếu hội thảo khoa học lần thứ 12 - Công nghệ thông tin và Ứng dụng trong các lĩnh vực (CITA 2023)

Linh Bui Khanh, Anh Nguyen Quynh, Hai Tran Van, Ha Nguyen Thi Thu 151

In this study, we used an LSTM model architecture in predicting customer attitudes
toward hotel services in Vietnam

Input gate: max length of sentences and use embedding

LSTM layer: 2

Dropout layer: 2

Activation: Softmax

5 Results

5.1 Data analysis

The study used data source from customer' online reviews that are generated on the
online booking TripAdvisor site. TripAdvisor is a platform that provides reliable travel
advice from travel customers. The TripAdvisor is considered the largest travel
community in the world, with over 200 million monthly visits to the site, nearly 100
million market members, and over 500 reviews from these travelers. The website is
used by many countries, about 80 countries around the world have used the website.
The TripAdvisor website is also a place where travel companies and guest houses check
all reviews from customers - those who have stayed at the hotel or used to participate
in their travel services. The study collects online customer reviews from TripAdvisor
about 12 hotel with 3-5 level stars in Vietnam, the list of hotels used for data collection
is as follows:

Table 1. List of hotels for data crawling
No Name of Hotel Star level
1. Thang Loi Hotel 4
2. Hanoi Lotte Hotel 5
3. Bao Son Hotel 4
4. Danang Intercontinental Hotel 5
5. Rex Hotel 5
6. Star City Nhatrang Hotel 3
7. Imperial Hue Hotel 3
8. Melia Hanoi 5
9. Hanoi Deawoo Hotel 5
10. Hanoi Hilton Opera Hotel 5
11. Fortuna Hotel 4
12. Intercontinental NhaTrang 5

WebHavy tool were used to collect data and stored as .csv file with these information:
Name of reviewer, date of review, content of review, title of review. Then, we using

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