基於虛擬世界購物喜好分析之個人化虛擬商品推薦

Abstract

隨著資訊技術的快速進步，虛擬世界特有的隱匿性讓使用者與他人可以更自在，在這樣的環境下促成虛擬世界愈來愈蓬勃發展。虛擬世界包含了虛擬生活、遊戲、教育性並提供虛擬社交互動功能，讓使用者進行互動建立虛擬社交生活圈。然而，虛擬世界的持續成長使得虛擬世界中的虛擬商品數量亦快速增加，使用者要在虛擬世界找到喜好的商品變得更加困難。 在虛擬世界中，使用者處於一個視覺化的生活環境，這種特有的視覺化商品組合，讓虛擬世界的使用者產生了獨特的體驗型購物，使得使用者的購物行為，受到一連串商品不同擺設的視覺化效果影響，同時同一商品組合亦會吸引使用者購買及蒐集。 本研究結合商品特性與商品組來考量使用者、社交生活圈鄰居及受訪者對虛擬商品之喜好，根據使用者之虛擬生活特性、社交互動與購物體驗之因素，設計新穎的虛擬世界使用者購物喜好分析，並以此為虛擬商品推薦機制。本研究實驗結果顯示，以虛擬世界使用者購物喜好分析為基礎的推薦方法，能改善虛擬商品推薦的準確性。

With the rapid development of information technology, virtual worlds form a competitive market due to its prosperous future. Users may manage their virtual life, play games and complete series of education courses in virtual worlds. Those activities all involve social interactions and induce the user to develop his/her social life circle. However, more and more virtual products are hastily provided to fulfill users’ various requirements, it becomes difficult to find virtual products suited for their virtual life. Hence, a personalized virtual products recommendation method fit for virtual worlds is demanded. A virtual world is a place full of visualized combination of virtual products. Users may visit others places and decorate his/her unique place with a variety of products. The unique visualized feeling while visiting others’ places gives rise a special shopping experience. Users’ buying behavior in the virtual world are also influenced by the effect of a series of virtual products and users prefer to collect virtual products belongs to the same item group. A novel virtual products recommendation approach is developed herein to recommend products based on the user and social circle’s preference, user’s shopping experience and product’s item group. The result shows that the proposed approach performs better performance than other virtual products recommendation methods.