.png){kind=icon}
.png){kind=icon}
Return
Generating method and application of basic probability assignment based on interval number distance and model reliability
.png){kind=icon}
PRE
AI
DOI：10.1007/s00500-023-09325-z.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
In the Dempster-Shafer (D-S) evidence theory, how to transform the objective data in reality into the basic probability assignment (BPA) is still an open issue. Based on this problem, a new method of generating BPA based on interval number distance model and reliability is proposed. First, construct the interval number model under each attribute. Second, calculate the interval number distance between the test sample and the interval number model and convert it into the initial basic probability assignment (IBPA). Thirdly, the final BPA is obtained by discounting the IBPA by constructing the comprehensive reliability from the static reliability and dynamic reliability of the interval number model. Finally, the Dempster combination rule is used to fuse the final BPA one by one, and the decision is made according to the fusion result. The ten-fold cross-validation results show that the classification accuracy under the three data sets is higher than other methods, and the classification accuracy of the Iris data set is 0.9733. At the same time, it is verified that the proposed method still has good effectiveness and robustness in the incomplete information environment.
Keywords:
Dempster-Shafer evidence theory
Basic probability assignment
Interval number distance
Comprehensive reliability