Home // DATA ANALYTICS 2019, The Eighth International Conference on Data Analytics // View article

A Multi-source Experimental Data Fusion Evaluation Method Based on Bayesian Method and Evidence Theory

Authors:
Huan Zhang
Wei Li
Ping Ma
Ming Yang

Keywords: data fusion; performance evaluation; Bayesian method; evidence theory

Abstract:
The experimental data used for system performance evaluation have many sources and multiple granularity. Thus, it is necessary to fuse multi-source experimental data for evaluation. Aiming at multi-source experimental data fusion problem, a multi-source experimental data fusion evaluation method based on Bayesian and evidence theory is proposed. According to the size of the data sample size, the large sample data are fused by the classical frequency method, while the small sample data are fused by Bayesian method. Then the parameter information is updated by the Bayesian method and data fusion result is obtained. The experimental data of different scenarios are fused by evidence theory. The evidence combination method is used to fuse the data when the evidence bodies do not conflict, while the weighted average correction method is used for data fusion when there is conflict between the evidence bodies. The result of the multi-source experimental data fusion is obtained based on Bayesian method and evidence theory.

Pages: 38 to 42

Copyright: Copyright (c) IARIA, 2019

Publication date: September 22, 2019

Published in: conference

ISSN: 2308-4464

ISBN: 978-1-61208-741-2

Location: Porto, Portugal

Dates: from September 22, 2019 to September 26, 2019