A Non-Parametric Method to Determine Basic Probability Assignment Based on Kernel Density Estimation

Abstract

Dempster–Shafer evidence theory has been extensively applied in a variety of fields due to its ability to solve knowledge reasoning and decision-making problem under uncertain environments. Nevertheless, it is still an open issue about how to determine the basic probability assignment (BPA). In this paper, a new non-parametric method based on kernel density estimation is proposed to determine BPA. First, the probability density function of each attribute is calculated, which can be regarded as the probability model for the related attribute using the training sample. Then, a nested BPA function is constructed using the intersections point of test sample and probability models. Finally, Dempster’s combination rule is used to combine multiple BPAs to get the final BPA. Some classification experiments are conducted on several datasets. The experimental results demonstrate that the proposed method is more effective and reasonable in determining BPAs, which has a better classification performance than the existing method.

Cite this work

No public copy is stored for this record, so only the metadata is cited.

Stored only when both papers are in the corpus — 43 citations exist upstream.

Papers that cite the same references — the relational fallback while embeddings are still backfilling.