雖然許多學者已經根據社群網站上使用者產生的內容所推斷的個人興趣，發展出各式各樣的推薦系統，極少學者體認到可以利用這些社群媒體獨有的特徵，透過機器人、使用者，和使用者的朋友之間的互動，推薦產品。本研究設計一支機器人，處理這個研究缺口，除了推薦產品給目標使用者，也推薦給他們的朋友圈。實驗結果證實這個推薦引擎的績效優於傳統的推薦機制。此外，本研究也提出假說，並且證實使用者的行為強度對於推薦效果產生顯著的影響。尤其是，經常張貼長訊息和獲得更多回應的活躍型使用者，和比較不活躍的使用者比起來，對推薦效果產生更大的影響。;Although researchers have proposed various recommendation systems based on the inferred interests provided by user-generated content on social networking sites, few researchers have realized that recommendations can take advantage of the characteristics of these social media in the form of interactions among bots, users, and users’ friend circles. This study designed a bot to address this research gap and recommended items to target users as well as their circle of friends. The experimental results confirmed that the recommendation engine outperformed other conventional recommendation mechanisms. Additionally, this paper also posits and confirms that user behavior intensity has a significant impact on recommendation effectiveness. In particular, active users who frequently post long messages and elicit more responses exerted a greater impact on recommendation effectiveness than less active users.
