「六子棋棋力」與「學科成績」相關性之研究

本論文研製之宗旨為：探討棋藝能力與學科能力之相關性。六子棋是個策略性人工智慧遊戲，具備「規則簡單」、「變化複雜」、「遊戲公平」之特性，由於規則簡單、易於入手，故受測者無須先備知識，僅需了解遊戲規則即可立即受測，然而受測者若想贏棋，落子對奕時，卻需具備一定程度的觀察、思考、推斷能力。本研究利用六子棋人工智慧遊戲與學生們對弈，取得棋力評估結果；並以五學科(國文英文數學自然社會) 六次成就評量平均，檢視學科成績；再用「亂度基礎分類法」、「粗糙集概念」篩選出重要屬性，並以「粒子群優化聚類」分析相關性，檢視六子棋棋力與五大考科學習成績之相關性，以做為棋藝能力暨學科能力相關性之評估參考。經研究分析：以KPSO驗證後有高達92.5%之正確率；以EBC篩選後正確率達82.5%；而用RSES簡化後正確率達90%，由此得知五大學科成績與六子棋棋力有高度相關性，且國文和數學與六子棋棋力有較高之相關性。

關鍵字

六子棋；學科成績；能力指標；亂度基礎分類法；粗集合理論；粒子群優化聚類

並列摘要

The goal of this study is to understand the relations between chess-abilities and school grades. Connect6 is an artificial intelligent game which is used to perform to be rule-simple, complicated variation and fair of competition. Based on the above consideration, the tester is very fair to undergo this game which is not required any pre-training. The game needs many techniques such as: observation, thinking and interference. We collected the data of the average on Chinese, English, Math, Nature, and Society to find the corresponding records of his/her chess ability. Then, the study used Entropy-Based Classification (EBC) and Rough Set Explanatory System (RSES) to search which courses are the governing factors. Finally, the Particle swarm optimization on Clustering technique is adopted the chess ability with considering different data groups of course combinations. The outcomes can offer some indices of chess-abilities and school grades. As the analyzed results, we obtain that the accuracy rate of original data through KPSO is 92.5%; accuracy rate through EBC is 82.5%; and accuracy rate through RSES is 90%. This outcome can make a conclusion for school grades has relations for chess ability. The data shows that Math and Chinese has higher relation for chess ability than other courses.