論文心得
1. 文章的Related Work部分值得借鑑
2. 文章的中心:
The main goal of this work is to develop a model for identification of the user’s interests for a Web search engine, using the user interactions stored in the query log files of the system.
3. 文章採用的兩個算法:
SVM 結合Error-Correcting Output Coding 來處理多類問題
PLSA(Probabilistic Latent Semantic Analysis)
4. 把query量化成vector 的方法
Each query was represented as a vector of terms that appeared in the documents giving an answer to the query |