學者指出現有的推薦系統與演算法具有共同的冷啟動(Cold Start)缺點,意即在一開始使用者尚未做出任何行為以前,系統無法得知該使用者的個人偏好及品味,此時儲存著巨量個人資訊的社群網站就能夠彌補此缺點,並且以協同式過濾的方法融合進現有的模式。本研究承襲上述學者的觀點,將社群元素納入個人化推薦系統中,研究中使用了消費者生活型態、消費者偏好建構協同式過濾規則,並以消費者社交互動導向、自我建構人格等指標成功將社交虛擬產品經驗的個人化推薦系統實體化。 研究結果發現,相對於一般電子商務網站使用的個人化推薦系統乃至於一般的產品介紹網站,納入了社交虛擬產品經驗後的個人化推薦系統,更能夠有效的協助消費者從大量的產品選擇中分辨出其可能感興趣的內容,而社群元素的導入也能夠有效提昇消費者滿意度與推薦系統整體品質。
Scholars have indicated that a common drawback called “cold start” for most present recommendation systems. By “cold start” means that before any activities or attempts made by the consumers, recommendation systems has no means to collect personal preferences or tastes. Social websites, storing vast amounts of personal data, can be a source to make up for the shortcomings of current recommendation systems, and can use a method called “Collaboration Filtering” to accommodate current recommendation algorithm. This study follows previous scholars’ efforts, trying to accommodate social elements into recommendation systems, using consumer lifestyle, consumer preferences to create collaboration filtering rules, and using indicators as social interaction orientation, self-constral successfully construct a social virtual product experience based recommendation system. As experiment result shows that, in comparison to the common recommendation systems used in e-commerce websites and even the product descriptive websites, recommendation system built with social virtual product experience not only helps consumers find out products they may be interested in more easily, but also helps to improve overall consumer satisfaction and recommendation quality more.