基於使用者經驗之多準則評分遊戲推薦系統

A Multi-criteria Game Recommendation System Based on User Experience

陳亭光(Ting-Kuang Chen)

指導教授 ： 曹承礎

國立臺灣大學/管理學院/資訊管理學研究所/碩士(2008年)

https://doi.org/10.6342/NTU.2008.01857

摘要

自推薦系統發展以來，隨其技術之進步，已廣泛應用於多種電子商務及個人活動之領域，包含電影、新聞、書籍、旅遊、飲食、休閒活動等。然而，電子遊戲為推薦系統尚未觸及之新興娛樂，僅視為相同於一般商品推薦可延伸之服務，但玩遊戲強調於使用者主動參與、個人技能養成、社群互動，產生沉浸效果，每位使用者所重視的因素皆不相同，因此使用者與遊戲產品的互動更為複雜化。
　　在使用者與產品、服務互動或從事特定活動的過程中，個人接受到的經驗亦對最終的喜好及滿意程度產生影響。當個人給予一項產品或服務負面評價時，可能僅是對過程中的特定經驗感到不快，不願再次接觸相同的經驗，而並非不滿意或排斥項目本身，但此現象無法在現有以單一項目評分為基礎的推薦系統中反映出來。
　　本研究採納依沉浸理論發展的遊戲沉浸模型，設計為衡量使用者經驗的準則，界定出適於遊戲推薦系統中衡量經驗之標準，作為多準則評分推薦系統的基礎，並提出運用各項經驗準則與總評分之間的相似度，視為調整使用者對各經驗準則重要性之權重，改良推薦預測的結果，應用於實作之遊戲推薦系統，並依實驗結果的數據分析與使用者回饋問卷兩方面，衡量出基於使用者經驗的推薦結果，更能貼近使用者的喜好與需求。本研究主要指標評估結果F1準確率值達86.25%，使用者對系統推薦的整體滿意度亦達73%。

關鍵字

推薦系統 ； 電子遊戲 ； 多準則評分 ； 協同過濾 ； 使用者經驗

並列摘要

Recommendation systems have been widely used for e-commerce and personal activities, including films, news, books, traveling and restaurants. Nevertheless, electronic games are new recreations which recommendation systems haven’t been applied to yet. Most people merely regard games recommendation as the extension of other commodities, while what really counts in playing games is active participation of users, cultivation of personal skill, interaction of communities, and immersion effects, hence the interaction between users and products are more complex.
Furthermore, personal experiences will certainly influence one’s preference and satisfaction during the period of user-product interaction. If a user gives a product or service a negative evaluation, one may merely feel unpleasant about some certain experience, and is unwilling to attain the same experience again. It doesn’t mean that the user is completely unsatisfied with the item, but this kind of situation cannot be reflected on the single criterion rating recommendation system.
This research adopts GameFlow Model which is based on Flow Theory to devise the criteria of user experience measurement, and use them for the rating elements of multi-criteria recommendation system. We adjust the criteria weight by analyzing the similarities between criteria and overall rating to enhance the result of prediction and prove the usefulness of our system. The primary measurements, F1-measure and users’ system satisfaction rate, reach to 86% and 73% respectively.

並列關鍵字

Recommendation System ； Electronic Games ； Multi-criteria Ratings ； Collaborative Filtering ； User Experience

參考文獻

[35] 陳建文，民94，電子遊戲產業分析與臺灣遊戲廠商之策略研究，國立台灣大學國際企業學研究所碩士論文

[1] Adomavicius, G., and Kwon, Y., “New Recommendation Techniques for Multicriteria Rating Systems,” IEEE Intelligent Systems, Vol. 22(3), pp. 48-55, 2007

[2] Adomavicius, G., and Tuzhilin, A., “Toward the Next Generation of Recommender Systems: A Survey of the State-of-the-Art and Possible Extensions,” IEEE Transactions on Knowledge and Data Engineering, Vol. 17(6), pp. 734-749, 2005.

[3] Apperley, T. H., “Genre and Game Studies: Toward a Critical Approach to Video Games Genres,” Simulation & Gaming, Vol. 37(1), pp. 6-23, 2006

[4] Balabnovic, B., and Shoham, Y., “Fab: Content-based, Collaborative Recommendation,” Communications of the ACM, Vol. 40(3), pp. 66-72, 1997.