基于支持度和虚假度迭代修正冲突证据

Iteratively Correcting Conflicting Evidence Based on Support and Falsity

摘要: 针对传统D-S组合规则在处理高度冲突证据时存在一定缺陷，难以获得合理的融合结果，提出基于新型的修正系数迭代修正冲突证据的方法。本方法通过兰氏距离计算证据之间的支持度，并将其与虚假度相结合，构建复合修正系数对证据进行预处理。接着，利用D-S组合规则对修正后的证据进行融合，并将所得结果作为新增证据，融入原始证据集实现信息强化，接着，再次计算新的修正系数，修正原始证据并继续使用D-S组合规则进行融合，对原始证据进行多次迭代修正，迭代的终止条件设定为连续两次融合结果的偏差小于设定阈值，在满足终止条件后，得到最终的融合结果。最后通过算例分析验证该方法在冲突证据的融合上具有可行性，与传统D-S组合规则相比，融合结果更加精确。

Abstract: The traditional Dempster-Shafer (D-S) combination rule has certain limitations when dealing with highly conflicting evidence, often failing to produce reasonable fusion results. To address this issue, this paper proposes a novel method for conflict evidence fusion based on an iterative correction mechanism using a newly developed modification coefficient. In this method, the Lance distance is first employed to calculate the support degree between pieces of evidence, which is then combined with the degree of falsity to construct a composite correction coefficient for evidence preprocessing. Subsequently, the D-S combination rule is applied to fuse the corrected evidence, and the resulting fusion outcome is incorporated into the original evidence set as new evidence to enhance the available information. This process is repeated iteratively: new correction coefficients are computed, the original evidence is re-corrected, and fusion is performed again using the D-S rule. The iteration terminates when the difference between two consecutive fusion results falls below a predefined threshold, at which point the final fusion result is obtained. Numerical examples are provided to verify the feasibility of the proposed method in fusing conflicting evidence. Compared with the traditional D-S combination rule, the proposed approach yields more accurate and reliable fusion results.