以情境感知資料與社群資訊建構餐廳推薦系統之研究
No Thumbnail Available
Date
2016
Authors
Journal Title
Journal ISSN
Volume Title
Publisher
Abstract
目前市面上大多的餐廳推薦系統當中所給予使用者的推薦資訊，皆透過開發者收集(網路相關資料、部落客體驗文章)，或經由系統平台上的使用者共同分享上傳而成的資料集。但這些資料會隨著時間長久的累積，亦或是開發人員對於資料沒有定期更新或維護，即造成了資料集過於老舊以及維護需求人力成本的問題。
本研究旨在藉由Facebook社群網站平台所提供的Graph API查詢語言技術，擷取其上的社群粉絲專頁餐廳資訊當作資料集，建構一套與Facebook同步的即時動態更新的餐廳推薦系統。此外本研究也結合情境感知(Context-aware)方式來開發應用，讓系統服務更能貼近使用者，在情境資料上利用使用者情境資料(Facebook帳號資訊、偏好餐廳設定)及實體情境資料(時間、地理位置)和使用者的Google Calendar事件；而推薦功能部分，本系統透過使用者對於個別餐廳頁面的評分紀錄，作推薦過濾的排序演算，其中也用到Facebook的按讚數，當作計算因素的依據，讓推薦達到更有效的過濾。最後，本系統實作採用響應式網站設計來建構一個可以在電腦、手機及平板都具有良好瀏覽效果的網站平台雛型。
Most traditional restaurant recommendation systems usually gather recommend information (i.e. internet source or blog article) from developers or members shared and uploaded to the system platform. These dataset probably are old and with potential maintain problems, which are brought about mainly by the increase of the dataset or by the lacking regularly update and maintenance. The study is aimed to develop a recommendation system by using Facebook released Graph API query language to retrieve page information about restaurants. Our system can be synchronized with the real-time information in Facebook. Additionally, we develop context-aware functions to provide convenient services for users. The context information includes user context (Facebook of user profile and preference setup), physical context (location and time) and user Google Calendar events. In this system, we use user rating scores of the restaurant and Facebook Page “likes” amounts as the filtering results to rate the recommendation list of the system. Finally, we implement a prototype website using Responsive Web Design methodology to support good user experiences on our web page for all devices including desktops, tablets, and phones.
Most traditional restaurant recommendation systems usually gather recommend information (i.e. internet source or blog article) from developers or members shared and uploaded to the system platform. These dataset probably are old and with potential maintain problems, which are brought about mainly by the increase of the dataset or by the lacking regularly update and maintenance. The study is aimed to develop a recommendation system by using Facebook released Graph API query language to retrieve page information about restaurants. Our system can be synchronized with the real-time information in Facebook. Additionally, we develop context-aware functions to provide convenient services for users. The context information includes user context (Facebook of user profile and preference setup), physical context (location and time) and user Google Calendar events. In this system, we use user rating scores of the restaurant and Facebook Page “likes” amounts as the filtering results to rate the recommendation list of the system. Finally, we implement a prototype website using Responsive Web Design methodology to support good user experiences on our web page for all devices including desktops, tablets, and phones.
Description
Keywords
Facebook, Graph API, 情境感知, 推薦系統, Facebook, Graph API, Context-aware, Recommendation System