資訊科技與網際網路越來越進步,使得網路上出現大量的資料及訊息,使用者要從中選擇所需的資料是不易的,加上智慧型行動裝置已成為人們生活中的必需品,人們開始透過智慧型行動裝置從事各式各樣的活動,因此如何迅速且準確的達成使用者的目標,已經成為在行動應用中很重要的課題,所以市場上逐漸出現許多推薦系統來協助使用者,近幾年情境感知技術與社群媒體被廣泛的運用,像是適地化服務、社群網站等應用,本研究利用無線射頻辨識技術與資料探勘建置社群分享推薦,利用無線射頻技術 RSSI 進行室內定位找出消費者的位置,再透過使用者的消費資訊進行資料探勘,找出消費者的消費習慣,消費者可以透過系統分享過去購買的商品到社群網站上,系統會截取使用者的社群脈絡,找出與使用者關系較親密的對象,推薦該對象所分享的商品給消費者,本研究發放模擬問卷方式收取 150 筆使用者的消費紀錄與相關資訊,使用消費紀錄進行 Apriori 資料探勘,並利用社群網站 Facebook 進行測試,透過 Facebook Friend Ranking 找出消費者最常接觸的對象,提供該對象所推薦的商品,並利用 Apriori 所找出的商品規則進行額外的推薦,透過此方法可以提升使用者購買意願,並提升其他商品的曝光度。
Information technology and internet are more and more advanced; a great amount of data and information appear on the internet and making users hard to select the information they need from it. Moreover, intelligent portable devices has become the necessity in life, and people start to conduct all kinds of activities via intelligent portable devices; thus, how to achieve the users’ goal rapidly and accurately has become a vital issue in mobile application, and many recommender systems have launched on the market to assist users. In recent years, context awareness technology and social media have been extensively applied to services such as location-based service and social websites. This research built community sharing and recommendation with RFID (Radio Frequency Identification) and data mining, using RSSI (Radio Signal Strength Indication) to localize the consumers indoor, and then find their purchasing habit by conducting data mining via the purchasing information of users. The users can share the purchased merchandise on the social websites via the system, and the system will retrieve users’ social networking, find objects that have closer relationship with the users and recommend products they bought to users. The research collected 150 purchasing records and related information of users via simulating questionnaires, using purchasing records to conduct Apriori data mining. The test was also conducted with social media Facebook to find objects who were contact the most with consumers via Facebook Friend Ranking, provide merchandise recommended by the objects, and use merchandise rules which was found with Apriori to recommend extra objects. This method may elevate users’ purchase intention and the exposure rate of other merchandise.