การวิเคราะห์ข้อมูลสะท้อนกลับด้านธุรกิจโรงแรม ผ่านบทวิจารณ์ออนไลน์ โดยใช้เทคนิคการวิเคราะห์ความรู้สึก

Abstract

The purpose of this research was to analyze the relationship between tourists' behavior and the sentiment analysis of online reviews, including categorizing reviews and words. The information is collected from the Agoda website, selecting only hotel information in 3 provinces that are popular for tourism and participating in the project “Rao tiew duay kun” including Phuket, Krabi, Phang Nga. There are 246,532 reviews. The analysis process is divided into 2 main processes. The first, process was conducted using Association rule. The second, process is conducting the sentiment analysis from online reviews. The first process aims to find the tourists' behavior and to compare with the Thai tourists behavior who participating in the project “Rao tiew duay kun”. There are two steps. The first step is to analyze the hotel stay behavior of tourists in many perspectives. They were compared using the Apriori and FP-Growth algorithms. The second step is a comparative analysis of the behavior of Thai tourists before the COVID-19 situation and during the COVID-19 situation. Both algorithms produce a total of 295 correlation rules. However, the execution time of FP-Growth, is 4.79 times faster than that of Apriori algorithm. In terms of sentiment analysis from online reviews, sentiment analysis was performed and the results were categorized (Topic Modeling) and display the results in Dashboard format.