JAIST Repository >
a. 知識科学研究科・知識科学系 >
a10. 学術雑誌論文等 >
a10-1. 雑誌掲載論文 >

このアイテムの引用には次の識別子を使用してください: http://hdl.handle.net/10119/9070

タイトル: Discounting and combination scheme in evidence theory for dealing with conflict in information fusion
著者: Huynh, Van-Nam
キーワード: Information fusion
Conflict management
Dempster-Shafer theory
Classifier combination
発行日: 2009
出版者: Springer
誌名: Lecture Notes in Computer Science
巻: 5861/2009
開始ページ: 217
終了ページ: 230
DOI: 10.1007/978-3-642-04820-3_20
抄録: Recently combination rules as well as the issue of conflict management in Dempster-Shafer theory have received considerable attention in information fusion research. Mostly these studies considered the combined mass assigned to the empty set as the conflict and have tried to provide alternatives to Dempster's rule of combination, which mainly differ in the way of how to manage the conflict. In this paper, we introduce a hybrid measure to judge the difference between two bodies of evidence as a basis for conflict analysis, and argue that using the combined mass assigned to the empty set emph{as a whole} to quantify conflict seems inappropriate. We then propose to use the discounting operator in association with the combination operator to resolve conflict when combining evidence, in which the discount rate of a basic probability assignment is defined using the entropy of its corresponding pignistic probability function. Finally, an application of this discounting and combination scheme to fusion of decisions in classifier combination is demonstrated.
Rights: This is the author-created version of Springer, Van-Nam Huynh, Lecture Notes in Computer Science, 5861/2009, 2009, 217-230. The original publication is available at www.springerlink.com, http://dx.doi.org/10.1007/978-3-642-04820-3_20
URI: http://hdl.handle.net/10119/9070
資料タイプ: author
出現コレクション:a10-1. 雑誌掲載論文 (Journal Articles)

このアイテムのファイル:

ファイル 記述 サイズ形式
14666-1.pdf174KbAdobe PDF見る/開く

当システムに保管されているアイテムはすべて著作権により保護されています。

 


お問い合わせ先 : 北陸先端科学技術大学院大学 研究推進課図書館情報係