多維度推薦系統：應用至行事曆助理

Multi-Dimension Recommender: The Application to Calendar Arrangement Assistant

隨著資訊科技的發展，帶來資訊產生資訊爆炸的現象。推薦系統提供了一個良好的解決方法，推薦系統透過分群與推薦的技術來達到減少資訊量與推估使用者潛在興趣的目的。目前推薦系統多半應用在單一維度的推薦，本研究則探討多維度推薦的應用，並以個人行事曆之助理軟體為例。助理軟體管理個人日常生活中的大小事務，成為最貼近個人的工具。而在個人行事曆的安排上，可以發現牽涉到有人、事、時、地、物五個維度。因此我們以五維度做基礎，透過合作推薦(Collaborative Recommender)的方式以達到個人潛在興趣的多維度(Multi-Dimensions)推薦，這將更能貼近使用者的需求，提供意想不到的資訊組合。

With the development of information technology, information overload has become a common phenomenon. Recommender systems could help relieve the information overload and provide users the most suitable or interesting products. Currently recommender systems only consider one dimension. This study would expand to multi-dimensions and choose the application of the calendar arrangement assistant software, which is a tool to manage all kinds of affairs in personal daily life. It could be found that five dimensions including people, thing, when, location and things, are involved in such arrangements. This study adopts cooperative recommendation approach to suggest multi-dimension recommendations, which would be more close to user's demands and provide more useful information mix.