請用此 Handle URI 來引用此文件:
http://tdr.lib.ntu.edu.tw/jspui/handle/123456789/59418
標題: | 樹狀抽樣式標註成本導向主動學習演算法 Annotation Cost-sensitive Active Learning by Tree Sampling |
作者: | Yu-Lin Tsou 鄒侑霖 |
指導教授: | 林軒田(Hsuan-Tien Lin) |
關鍵字: | 機器學習,主動學習,標註成本導向, Machine Learning,Active Learning,Annotation Cost-sensitive, |
出版年 : | 2017 |
學位: | 碩士 |
摘要: | Active learning is an important machine learning setup for reducing the labelling effort of humans. Although most existing works are based on a simple assumption that each labelling query has the same annotation cost, the assumption may not be realistic. That is, the annotation costs may actually vary between data instances. In addition, the costs may be unknown before making the query. Traditional active learning algorithms cannot deal with such a realistic scenario. In this work, we study annotation-cost-sensitive active learning algorithms, which need to estimate the utility and cost of each query simultaneously. We propose a novel algorithm, the cost-sensitive tree sampling(CSTS) algorithm, that conducts the two estimation tasks together and solve it with a tree-structured model motivated from hierarchical sampling, a famous algorithm for traditional active learning. By combining multiple tree-structured models, an extension of CSTS, the cost-sensitive forest sampling(CSFS) algorithm, is also proposed and discussed. Extensive experimental results using data sets with simulated and true annotation costs validate that the proposed methods are generally superior to other annotation cost-sensitive algorithms. |
URI: | http://tdr.lib.ntu.edu.tw/jspui/handle/123456789/59418 |
DOI: | 10.6342/NTU201701006 |
全文授權: | 有償授權 |
顯示於系所單位: | 資訊工程學系 |
文件中的檔案:
檔案 | 大小 | 格式 | |
---|---|---|---|
ntu-106-1.pdf 目前未授權公開取用 | 6.76 MB | Adobe PDF |
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。