推薦系統簡介

推薦系統的目的是向使用者提供有價值的項目(items)，例如商品、新聞和串流影片，以幫助他們在大量資訊中找到有價值的內容。我們通常研究的推薦系統是個人化的，根據個人或特定使用者群體的偏好調整推薦項目。相比之下，大眾化的推薦系統方法通常較容易實現，例如top10 熱門項目。儘管這種方法在特定情境下可能有效，但它並不是推薦系統主要討論的議題。個人化推薦最簡單的形式可以用一排序後的項目列表呈現(最有興趣->最沒興趣)，系統透過使用者的偏好(對某項目的評分、是否查看某項目的詳細資訊)或某些條件排序。最早的推薦系統透過尋找和被推薦使用者(目標)相似的使用者，並將相似使用者關注或選擇的項目推薦給目標，該方法即為協同過濾(Collaborative filtering)。而現代人日常要接受的資訊繁多2，如何令使用者可以方便的取得有價值的資訊已成為重要問題，推薦系統很好的解決了此問題。

更新計畫與週期

可能兩周一更

本書架構

Part I Basic Techniques

  1. Data Mining Methods for Recommender Systems 🚧
  2. Content-based Recommender Systems: State of the Art and Trends 🚧
  3. A Comprehensive Survey of Neighborhood-based Recommendation Methods 🚧
  4. Advances in Collaborative Filtering 🚧
  5. Developing Constraint-based Recommenders 🚧
  6. Context-Aware Recommender Systems 🚧

Part II Applications and Evaluation of RSs

  1. Evaluating Recommendation Systems 🚧
  2. A Recommender System for an IPTV Service Provider: a Real Large-Scale Production Environment 🚧
  3. How to Get the Recommender Out of the Lab? 🚧
  4. Matching Recommendation Technologies and Domains 🚧
  5. Recommender Systems in Technology Enhanced Learning 🚧

Part III Interacting with Recommender Systems

  1. On the Evolution of Critiquing Recommenders 🚧
  2. Creating More Credible and Persuasive Recommender Systems: The Influence of Source Characteristics on Recommender System Evaluations 🚧
  3. Designing and Evaluating Explanations for Recommender Systems 🚧
  4. Usability Guidelines for Product Recommenders Based on Example Critiquing Research 🚧
  5. Map Based Visualization of Product Catalogs 🚧

Part IV Recommender Systems and Communities

  1. Communities, Collaboration, and Recommender Systems in Personalized Web Search 🚧
  2. Social Tagging Recommender Systems 🚧
  3. Trust and Recommendations 🚧
  4. Group Recommender Systems: Combining Individual Models 🚧

Part V Advanced Algorithms

  1. Aggregation of Preferences in Recommender Systems 🚧
  2. Active Learning in Recommender Systems 🚧
  3. Multi-Criteria Recommender Systems 🚧
  4. Robust Collaborative Recommendation 🚧

reference

[1] Wiki: Collaborative filtering

[2] Wiki: Information overload