A Machine-Learning-Based Approach for Tourist Arrival Trend Prediction

Aug 3, 2026

Uploaded by Levi

This study proposes a machine-learning-based technique to predict trends in tourist arrivals using online news headlines and historical tourist-arrival counts. Forecasting tourist arrivals supports destination governments and tourism businesses in planning services and mitigating demand uncertainty. Models based on Logistic Regression and Support Vector Machine are trained with 47,298 Indonesian online news headlines. Results indicate Logistic Regression reaches up to 67.4% F-score, while SVM reaches up to 62.9% F-score.

LanguageEnglish

CategoryResearch & Report

FormatPDF

Pages22

Reading Time55 min

AccessFree to download

Introduction

Related Work

Methodology

Results and Analysis

Discussion

Conclusion and Future Work

Journal of Business on Hospitality and Tourism Vol. 9 Issue 2 (2023) 212-233 DOI: [https://doi.org/10.22334/jbhost.v9i2.480](https://doi.org/10.22334/jbhost.v9i2.480) [Online since: December 30](Online since: December 30), 2023 Submitted: 22/10/2022, Revised:22/10/2023, Accepted: 1/12/2023 [https://jbhost.org/](https://jbhost.org/) A Machine-Learning-Based Approach for TouristArrival Trend Prediction I Putu Edy Suardiyana Putra 􀀍 Department of Digital Business Institut Pariwisata dan Bisnis Internasional, Bali, Indonesia Denok Lestari Department of Digital Business Institut Pariwisata dan Bisnis Internasional, Bali, Indonesia Komang Ratih Tunjungsari University of Otago, Dunedin, New Zealand 􀀍 [edy.suardiyana@ipb-intl.ac.id](edy.suardiyana@ipb-intl.ac.id) Abstract This study proposes a machine-learning-based technique to predict trend in tourist arrivals based on online news headlines and the number of previous tourist arrivals. Tourist arrivals prediction is important to give information to destinations’ local governments and businesses to prepare their services. We use Logistic Regression and Support Vector Machine to create a model to predict the increase in tourist arrivals monthly. News headlines from three online Indonesian news portals are used. A total of 47,298 online news headlines were collected. The results show that Logistic Regression can achieve up to 67.4% ofF-score while Support Vector Machine can achieve up to 62.9% of F-score. These results show that adding online news headlinesand machine-learning algorithms can give significantly better results in predicting tourist arrivals. Keywords Tourist-Arrival Trend, Machine Learning, News Headlines, Logistic Regression, Support Vector Machine. © 2023 Authors. This work is licensed under a Creative Commons Attribution 4.0 International (CC BY 4.0) . All writings published in this journal are personal views of the authors and do not represent the views of this journal and the author's affiliated institutions. Introduction Tourism is one of the fastest growing industries in the world without a doubt. Forecasting tourism demand is important to reduce the risk of the tourism service vendors losing their business since the tourism service or product is short-lived. Although there are many studies have been conducted to show tourism demand, a common metric that is used to measure tourism demand is tourist arrival(Song & Li, 2008) . Forecasting tourist arrival is important to the destination’s local government and numerous tourism service vendors to develop their strategic planning so that they can get benefit from foreign exchange income and other potential economic benefits (Yuan, 2020) . Thus, predicting tourist arrivals is crucial for the tourism industry. Studies from(Li, 2022; Purnaningrum & Athoillah, 2021; Xie et al., 2021; N. Yu & Chen, 2022; Yuan, 2020), try to develop techniques to predict tourism demand using machine learning. Data from Twitter, the Baidu search engine, the Central Bureau of Statistics of Indonesia, and past tourism arrival data are used by these studies to train the machine learning algorithms. Their research shows that machine-learning-based algorithms are able to predict tourist arrival relatively well. Another study from(Park et al., 2021), shows that including online news data can improve the performance of the algorithm to predict tourist arrival. Park et al. also mention that although using data from online news can give a better result, this data source remains unexplored. In his study, shows that news readers usually scan the headlines, and rarely read the entire text (Dor, 2003) . Studies from(Liu et al., 2018; Oncharoen & Vateekul, 2018), show that using news headlines can produce relatively good results in predicting stock price. Therefore, this study uses online news headlines together with the number of previous tourist arrivals asthe main source of data to train the machine learning algorithm

FAQs

What data sources are used to predict tourist arrival trends?

The approach uses online news headlines from Indonesian news portals and the number of previous tourist arrivals as the main input for the model.

Which machine-learning models are evaluated in the study?

The study trains and evaluates Logistic Regression and Support Vector Machine to predict whether monthly tourist arrivals increase or decrease.

How effective are the models, and does adding news headlines help?

Logistic Regression achieves up to 67.4% F-score and SVM up to 62.9% F-score. Incorporating both historical arrivals and news headlines improves classifier performance, especially for recall and F-score when using Logistic Regression.

Aug 3, 2026

Uploaded by Levi