用戶數對推薦系統的影響

雖然東半球的我們在過春節，但是世界還是照常運轉的，Daniel Lemire 剛剛在他的部落格寫了一篇隨筆 - 需要多少用戶才能建構一個有效率的 collaborative filtering 系統？

大過年的，沒有太多心思想嚴肅的事情，所以隨手把「沒有深思熟慮」的想法加在Daniel 的想法後面：

• You can build an effective recommender system with as little as two people.
  
• 兩個用戶何必要什麼推薦系統？拳頭就是推薦系統
  
• As you have more users, you tend to have more training data. Hence, you may have more accurate recommendations.
• 還是那句老話：Worry about the data before you worry about the algorithm
• More accurate recommendations may not be important to your users.
  
• Yes, we've know that very well - Prediction accuracy is not enough，前一篇談如何評估推薦系統時，我已經提到要花點時間想想 usefulness 是怎麼一回事兒了。Stay tuned.
  
• The exact count of your users may not matter as much as the diversity of your users.
  
• 我完全同意， diversity 是影響推薦系統更關鍵的因素
  
• A good rule of thumb is that you should have many more users than you have items to recommend.
• Given the right algorithms, your accuracy will improve monotonically with the number of users and the amount of training data.
• The users may enter feedback data to correct the assumptions of your recommender system and thus, improve it over time.
• Ya! 用戶的回饋是 CF 最重要的"資產"了
  

最後還是要加個 Disclaimer 比較保險，還是請諸君看原文，而且不要忘了看精彩的讀者回應喔！

All in all, ，今日是戊子年大年初一，小弟在此向所有相識或不相識的您拜年，戊子大吉，新年快樂！