Data mining techniques for tourist review classification
A large amount of information has been provided by the increasing volume of user generated content, through social networking services like reviews, comments and past experiences. Online review has become one of the most influential information sources for consumer decision-making. This information...
Saved in:
| Main Author: | |
|---|---|
| Format: | Thesis |
| Language: | English |
| Published: |
2019
|
| Subjects: | |
| Online Access: | http://eprints.utm.my/id/eprint/96379/1/MustaphaAbubakarGiroMSC2019.pdf.pdf |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | A large amount of information has been provided by the increasing volume of user generated content, through social networking services like reviews, comments and past experiences. Online review has become one of the most influential information sources for consumer decision-making. This information is freely accessible online and used to support tourist decision-making process. Despite several studies conducted on tourist online reviews, there have been limited studies exploring tourist reviews’ ratings for 1 - 5 reviews star in predicting tourist response to an attraction. This study aims to predicting tourist ratings based on the tourist textual response (reviews) made on Petronas Twin Tower in Kuala Lumpur that is freely available on TripAdvisor. This is devised by building a predictive classification model that predicts the rating a tourist will possibly give. A qualitative approach is adopted where data miner tool was to collect tourist reviews from TripAdvisor; and the reviews dataset was preprocessed in Rapidminer to generate sentiment values which was fed to the models after some transformation. The sentiments gained/produced is utilized to compare which classification model gives the best prediction in terms of accuracy. The result showed that MLP prediction model returns a promising result in terms of accuracy over other techniques for predicting tourist response based on ratings (1-5) in which has 19% better accuracy than the other techniques tested. In conclusion, this study could contribute to the field of study by introducing a predictive model and could help destination marketers evaluate tourists’ responses to a certain destination in advance, and could also potentially influence the final destination choice by improving marketing strategies accordingly. Destinations might use these analyses to predict the weaknesses or strengths of their image based on the analysis of tourists’ reviews. |
|---|