# Temporal evidence fusion evaluation method considering time sequence variation trend

## TL;DR

This paper introduces a new method for evaluating temporal evidence by considering time sequence trends to improve fusion accuracy and reduce uncertainty.

## Contribution

A novel temporal evidence fusion method that integrates time sequence variation trends to handle conflicts and improve evaluation.

## Key findings

- The proposed method increases discrepancy between combined BPAs while reducing uncertainty.
- It effectively handles conflicting information in temporal sequences.
- Validation using health assessment data shows improved fusion results.

## Abstract

Aiming at the temporal information fusion evaluation problem, a temporal evidence fusion evaluation method was proposed by considering the time sequence trend based on evidence theory to fully reflect the dynamic trend of temporal information and the influence of the time factor on fusion evaluation. Firstly, the trend of temporal evidence sequence was integrated into fusion evaluation. The temporal variation factor of the proposition was defined by analyzing the sequence of changes in evidence to measure the dynamic of the proposition. Then, conflict information between evidence was interpreted as the temporal variation information of propositions. An evidence combination rule integrating temporal variation factors of propositions was proposed in this paper. Finally, the proposed method was used to fuse and evaluate the temporal evidence. The consecutive periodic health assessment data were utilized to validate the fusion results. Numerical examples show that the discrepancy between the combined BPAs obtained using the proposed method is significantly larger, while the associated uncertainty is notably reduced. The proposed method is capable of handling conflicting information within a temporal information sequence and obtaining a reasonable fusion evaluation result by combining the temporal variation trend of propositions. It provides an idea for information fusion evaluation considering the time sequence trend.