Improvement of the RF-based and XGBoost-based visitor data prediction for colocated museums

Name: Warissara Limpornchitwilai

Organization : Thammasat University. Sirindhorn International Institute of Technology

Name: Prapun Suksompong

Organization : Thammasat University. Sirindhorn International Institute of Technology

Name: Chalie Charoenlarpnopparut

Organization : Thammasat University. Sirindhorn International Institute of Technology

Name: La-or Kovavisaruch

Organization : Thailand Science Park. National Electronics and Computer Technology Center

Email : la-or.kovavisaruch@nectec.or.th

Name: Taweesak Sanpechuda

Organization : Thailand Science Park. National Electronics and Computer Technology Center

Email : taweesak.sanpechuda@nectec.or.th

Name: Nakatani, Momoko

Organization : Tokyo Institute of Technology. School of Engineering

Email : nakatani.m.ad@m.titech.ac.jp

keyword: Random forests (Machine learning)

LCSH: Museums -- Management

; XGBoost (Computer program language)

LCSH: Forecasting

LCSH: Machine learning

LCSH: Data mining

Abstract: Over recent years, data analysis and prediction have been crucial for museum curators to predict the number of visitors and plan how to improve their services. In this research, the factors affecting the number of visitors are investigated and analyzed. A museum visitor prediction method based on the Random Forest (RF) model and the XGBoost model is proposed using historical data from 2016 to 2020 collected from three museums under the supervision of the National Science Museum Thailand. Since all museums are located in the same area, the proposed method combines data from the three museums to improve prediction performance. More specifically, the result from one museum is used to help predict visitor numbers for the other museums. Originally, the RF and XGBoost models achieved average accuracy values of 59.14% and 58.08%, respectively. With the proposed method, the average accuracy values increase to 72.04% and 74.02%, respectively. The results also show that the XGBoost model performs better and consumes less computational time than the RF model. In addition, the XGBoost model achieves 96.77% accuracy after reducing the number of visitor classes from five to three

King Mongkut's University of Technology North Bangkok. Central Library

Address: BANGKOK

Email: library@kmutnb.ac.th

Created: 2022

Modified: 2026-05-12

Issued: 2026-05-12

บทความ/Article

application/pdf

BibliograpyCitation : In Electrical Engineering/Electronics, Computer, Telecommunications and Information Technology (ECTI) Association, Thailand. 2022 19th International Conference on Electrical Engineering/Electronics, Computer, Telecommunications and Information Technology (ECTI-CON 2022) (1570793614). Bangkok : ECTI Association

eng

©copyrights King Mongkut's University of Technology North Bangkok

ใช้เวลา

0.028623 วินาที