| 摘要: | 目的：本研究建構國際大型運動賽會國家參賽效率分析與預測模式。方法：經由次級資料分析，蒐集龐大數據、介接資料庫等作法，從「效率與生產力」、「趨勢分析」、進化到「預測」未來，開創完整的預測模式，探討亞洲運動會亞洲國家參賽效率管理模式、以及參賽效率預測分析。本研究運用效率分析衡量1990年北京亞洲運動會至2023年杭州亞洲運動會之亞洲國家歷屆參賽效率分析，同時運用倒傳遞類神經網絡資料包絡分析法預測2026年愛知名古屋亞洲運動會之亞洲國家競技成績。結果：研究顯示，2026年第20屆愛知名古屋亞洲運動會國家參賽效率，總獎牌預測結果上升：中國、日本、南韓、伊朗、泰國、馬來西亞、新加坡、蒙古等8個國家；總獎牌預測結果維持：中華臺北、卡達、菲律賓等3個國家；總獎牌預測結果下降：印度、中國香港、中國澳門等3個國家與地區。最終建構亞洲國家大型運動賽會績效評估系統。結論：本研究運用資料包絡分析法、類神經網路、ANNs結合DEA、時間序列等方法，包含管理科學、決策分析、人工智慧等不同層面之研究方法，切入當前國內運動AI人工智慧研究相對缺乏之缺口，對於國家體育運動政策與發展議題，連結「預測」國際運動賽會競技表現管理實務，提供國內運動主管單位政策制定之參考，達到學術與實務結合的貢獻與效益。
Purpose: This study aims to construct an analysis and prediction model for the national participation efficiency in Mega Sports Events. Methods: By analyzing secondary data and collecting extensive data through database integration, the study progresses from "efficiency and productivity" to "trend analysis" and ultimately to "prediction" of the future, creating a comprehensive prediction model. The study investigates the efficiency management model of Asian countries participating in the Asian Games and conducts a predictive analysis of their participation efficiency. We focus on analyzing the historical participation efficiency of Asian countries from the 1990 Beijing Asian Games to the 2023 Hangzhou Asian Games while also predicting the participation efficiency and athletic performance of Asian countries in the 2026 Aichi-Nagoya Asian Games. Results: The efficiency of national participation in the 20th Asian Games in Aichi-Nagoya in 2026 has shown an increase in the total medal prediction for eight countries: China, Japan, South Korea, Iran, Thailand, Malaysia, Singapore, and Mongolia. The total medal prediction has maintained for three countries: Chinese Taipei, Qatar, and the Philippines. However, the total medal prediction has decreased for three countries and regions: India, Hong Kong (China), and Macau (China). Eventually, we created the Asian nations' performance assessment system for mega sports events. Conclusion: This study utilizes Data Envelopment Analysis, Artificial Neural Networks, and time series analysis methods. It incorporates different research approaches from management science, decision analysis, and artificial intelligence. By addressing the current lack of research on sports AI and artificial intelligence in the domestic context, we aim to bridge the gap and contribute to national sports policies and development issues. It connects the "prediction" of Mega Sports Events' performance with management practices, providing a reference for domestic sports government policy-making. The study seeks to achieve the integration of academia and practice, offering contributions and benefits in the field. |
