在科技發達的今日,無所不在的情境感知服務(Context-Aware Services)逐漸的成為新的趨勢。然而,要提供個人化的推薦,勢必要有充分的資訊及良好的推論模型(Inference Models),因此,本研究提出應用情境感知及模糊理論於智慧型推薦系統,藉由情境感知(Context-Awareness)的概念來獲取使用者週遭的情境資訊(Contextual Information),接著透過模糊推論引擎(Fuzzy Inference Engine)將專家的經驗結合到推論當中,藉此達到個人化的推薦。 此外,本研究也以個人化的廣告推薦為例,從使用者的交談內容及個人資訊中來蒐集短期和長期的情境資訊,並且利用這些情境資訊經由模糊推論引擎找出符合使用者需求的廣告,藉此發展一個具有個人化廣告特色的交談平台。最後,由本論文的實驗結果顯示,本研究所出的個人化廣告推薦確實能夠改善傳統廣告的點閱率,如此將可讓廣告主獲得良好的廣告效益,也能讓使用者取得感興趣的廣告,最後創造使用者及廣告主雙贏的局面。
Nowadays, the information technology is developed by leaps and bound. The ubiquitous context-awareness services have become new trends. Both sufficient information and good inference model are very important to obtain an effective recommendation. So, we propose a smart recommender system which integrates the technologies of context-awareness and fuzzy set theory. We employ the context-awareness technology to perceive the contextual information around. And, take the advantages of fuzzy set theory to combine the experiences of experts for the inference engine. Thus, the smart personal recommendation can be achieved. In our experiment, we extract the conversation contents of users in a conversation platform to obtain short-term contextual information and use the personal information of users to be long-term contextual information. We employ these contextual information to find the advertisements by fuzzy inference system. Finally, through the experiment results which we can improve the click rate of advertisement. Therefore, the advertisers can acquire more effective advertising and the user can also acquire advertisements which he like.