Improving classification models with taxonomy information Presenter : Chang,Chun-Chih Authors : Luca Cagliero , Paolo Garza 2013 DKE
Outlines Motivation Objectives Methodology Experiments Conclusions Comments
Motivation The classification problem has extensively been investigated by the research community. A number of different approaches to build accurate classifiers have been proposed . Those approaches have problem of classifying structured data supplied with taxonomies. 監督學習的方法有些問題，最顯著的問題是他們為了準確的學習必須要有大量的標籤數量訓練，雖然未標籤文件是很簡單收集和豐富的，但標籤文件是很不好產生的，因為標籤任務一定要由人類發展完成
Objectives This paper presents a general-purpose strategy to improve structured data classifier accuracy by enriching data with semantics-based knowledge provided by a taxonomy built over data items. We also presented a generalized version associative classifier, namely the G−L3 classifier, which makes use of taxonomies to drive classifier learning.
Methodology
Methodology
Experiments-
Experiments-
Experiments
Experiments
Experiments
Experiments
Experiments
Conclusions Experiments show the effectiveness of the proposed approach in improving the accuracy of state-of-art classifiers, associative and not. .
Comments Advantages Applications - Taxonomies - Classification The proposed approach could directly access or easily infer meaningful taxonomy models over the analyzed data. Applications - Taxonomies - Classification