虛擬世界之虛擬商品推薦

Abstract

隨著網際網路的興起，網路遊戲也愈來越多，而虛擬世界就在這樣的趨勢下蓬勃發展。然而，隨著虛擬世界持續的成長，虛擬世界中的虛擬物品數量也越來越多，資訊過載的問題使得使用者對於在虛擬世界上找到喜好的物品變得更加困難。為了解決資訊過載的問題，推薦系統便是有效解決問題的方法。然而，現有的推薦系統並未分析虛擬世界的特性，使用者在虛擬世界經營虛擬生活會有獨特的視覺化商品組合；在虛擬世界與他人互動也與現實世界有所差異。因此分析虛擬世界的特性來設系推薦系統是重要的。 本研究在虛擬商品推薦之研究將考量使用者虛擬生活特性與社交互動之因素，分析社交生活圈鄰居與互動模式對於使用者購買虛擬商品之影響，並設計符合虛擬世界使用者喜好分析以及虛擬商品推薦機制。本研究蒐集 Roomi 虛擬世界網站資料進行實驗評估，針對使用者的社交互動與購買紀錄進行分析。本研究實驗結果顯示，利用虛擬世界的虛擬生活特性與社交互動因素為基礎的推薦方法能改善推薦的準確性。

With the rapid development of Internet, more and more online games are attached to famous websites. With this trend, virtual worlds are increasingly popular and rapidly getting more attention. However, with the growing trend of virtual worlds, more and more virtual goods are made available, making it hard for users to find objects suited for their virtual life. This illustrates the importance of developing a personal virtual goods recommendation system based on the characteristics of virtual worlds to solve this information overload problem which has not been explored in the past. In this work, we propose a novel recommendation method based on the characteristics in users’ virtual lives and the contact influence from the social interaction to recommend virtual goods on a virtual world website - Roomi. We collected the dataset from Roomi to evaluate our proposed approach. In order to evaluate our method, we collected the historic buying data and the social interactions on this platform. Our experiment results show that our proposed methods can improve the prediction accuracy and the quality of recommendation.