https://scholars.lib.ntu.edu.tw/handle/123456789/149804
標題: | 預測網路共享資源之人為標籤 Prediction of social annotation on resource-sharing services |
作者: | 陳如芃 Chen, Ju-Peng |
關鍵字: | 標籤,資料檢索;folksonomy;information retrieval;algorithm;social resource sharing | 公開日期: | 2007 | 摘要: | 預測網路標籤趨勢 Popular social bookmark service del.icio.us enables easy annotation for user to orga- nize their resources. The lightweight conceptual structure built by users called folkson- omy is able to provide a di?erent retrieval service that utilzes its power. However, the performance was hindered by lack of tags at a resource’s new arrival. Thus, our work aims to overcome the handicap of retreival for new-coming URLs by predicting tags at early stage. We exploit accumulated tagging records from users to predict tags. Our experiements on del.icio.us URLs show that our algorithm has a high coverage of tags appearing in the mature stage. Our prediction captures 80% of a 13-month old tag set at the first month and 80.23% of 100-users tag set with 5-users tag set. |
URI: | http://ntur.lib.ntu.edu.tw//handle/246246/53517 | 其他識別: | en-US |
顯示於: | 電機工程學系 |
檔案 | 描述 | 大小 | 格式 | |
---|---|---|---|---|
ntu-96-R94921033-1.pdf | 23.31 kB | Adobe PDF | 檢視/開啟 |
在 IR 系統中的文件,除了特別指名其著作權條款之外,均受到著作權保護,並且保留所有的權利。