透過您的圖書館登入
IP:18.119.111.9
  • 會議論文
  • OpenAccess

結合鄰近感知之社群網路朋友推薦系統之研製

摘要


現有的Facebook好友推薦系統是推薦使用者的同學,及同學的同學,層層的關係去推薦使成為好友,但常常你看到的推薦名單上的人是你完全不認識的陌生人,所以成功加為好友的機率很低。為此我們設計一個推薦系統,利用智慧型手機上低功耗藍牙(Bluetooth LE, BLE)來傳接封包及利用Facebook上的用戶之間的共同好友資訊當作親密度的資料來源,進行計算後得到使用者與其他用戶的個別親密度。本論文提出使用鄰近感知技術藍牙來偵測有哪些用戶在身邊,因此推薦名單上出現的用戶將會是使用者日常生活中真實遇到的人,等同於是近距離的交友。而且兩人之間友共同好友,使用者可以向共同好友詢問用戶的更多詳細資訊,可以確保交友的安全性,避免網路惡意詐騙。最後,本論文實作鄰近感知朋友推薦系統與實驗。

並列摘要


Friend Recommendation of Facebook is to recommend user's classmates, and classmates' classmates to make friends. But often people at recommended list is the stranger to user, the probability of succeeded making friends will be very low. To increase probability, we design a recommended system that uses Bluetooth Low Energy (Bluetooth LE, BLE) in smartphone to receive and sent package in terms of using common friend's information at Facebook to calculate affinity of each pair of users. Because of using BLE to detect how many other users nearby user, people at recommended list will be the person who user may encounter at daily life. That equal to making friends at close distance. Moreover, user can ask common friend to know more information about new friend, this can ensure safety of making friends and avoid Internet fraud. Finally, the proposed friend recommendation system is designed and implemented.

延伸閱讀