基於多特徵探勘的電腦遊戲個人化推薦 = Personalized Rec...
國立高雄大學資訊工程學系碩士班

 

  • 基於多特徵探勘的電腦遊戲個人化推薦 = Personalized Recommendation of PC Games Based on Multiple Features Mining
  • 紀錄類型: 書目-語言資料,印刷品 : 單行本
    並列題名: Personalized Recommendation of PC Games Based on Multiple Features Mining
    作者: 許聰榮,
    其他團體作者: 國立高雄大學
    出版地: [高雄市]
    出版者: 撰者;
    出版年: 2012[民101]
    面頁冊數: 54面圖,表格 : 30公分;
    標題: 推薦系統
    標題: Recommendation System
    電子資源: http://handle.ncl.edu.tw/11296/ndltd/82724319500039967523
    附註: 104年10月31日公開
    附註: 參考書目:面45-46
    摘要註: 近年來,隨著網際網路的發達以及遊戲開發技術的成熟,造就電腦遊戲市場的快速成長,玩家在面對琳琅滿目的電腦遊戲當中,不容易選擇適合玩家所喜好的遊戲。正因為如此,如何幫助玩家選擇適合、感興趣、相關性高的遊戲來推薦給玩家,成為一個重要的研究議題。在本篇論文中,我們提出一個有效的推薦策略,採用基於Apriori加權的資料探勘方法,從玩家的遊戲歷史記錄,找出不同群體玩家共通的遊戲特徵項目,作為多特徵推薦系統的基礎,並應用CAST分群的方法,找出玩家之間的相似度,再應用PageRank演算法,推算出玩家之間共同喜好的遊戲排名清單,並依據個體玩家的喜好,匹配相似的群體玩家,採用群體玩家的喜好遊戲排名清單,來協助玩家更容易找到感興趣的遊戲。本研究中所提出之推薦策略,應用於實作之遊戲推薦系統,並依實驗結果的資料分析,衡量出基於多特徵探勘的推薦結果。實驗結果顯示,所提出策略的有效性。 With rapid in Internet and game-playing technologies, the computer games market has seen explosive growth in recent years. However, it is often not easy for a player to find a game that meets their preferences. In this paper, we propose a recommendation strategy based on a player’s game-playing history, and use the Apriori-based weighted data mining method to identify common features of games among different groups of players. The recommendation system considers multiple features and adopts the CAST clustering method to assess the similarities among players. It then modifies the PageRank algorithm to calculate the different groups of players in the game ranked list. The system then matches the game ranked list of the groups of similar players, and thus helps players to find games which they are more likely to be interested in. The recommendation strategy proposed in this thesis is also used for a real implementation of a game recommendation system, and experiments are carried out. The results show the effectiveness of the proposed strategy.
評論
Export
取書館別
 
 
變更密碼
登入