Evidence theory is widely used in data mining, machine learning, clustering and database systems. In these applications, often combination of mass functions is performed without checking the degree of consistency between the mass functions, which may lead to counterintuitive results. In this paper, we aim to measure the divergences among mass functions which can hence prevent highly inconsistent mass functions from been combined. To this end, we propose a divergence measure between two mass functions. In addition, incompleteness measures and similarity measures are also provided based on divergence measures.
|Title of host publication||Knowledge Science, Engineering and Management: 8th International Conference, Chongqing, China, October 28-30 2015, Proceedings|
|Editors||Songmao Zhang, Martin Wirsing, Zili Zhang|
|Place of Publication||Switzerland|
|ISBN (Print)||Online: 978-3-319-25159-2, Print: 978-3-319-25158-5|
|Publication status||Published - 3 Nov 2015|
Bibliographical noteThere is no full text available.
Ma, J. (2015). A Divergence Measure Between Mass Functions. In S. Zhang, M. Wirsing, & Z. Zhang (Eds.), Knowledge Science, Engineering and Management: 8th International Conference, Chongqing, China, October 28-30 2015, Proceedings (Vol. 9403, pp. 53-65). Switzerland: Springer Verlag. https://doi.org/10.1007/978-3-319-25159-2_5