基于证据推理的相邻碎片判定方法

摘要

Aiming at the uncertain fuzzy error matching problem in the feature information of cultural relic's fragments,this paper proposed a multi feature matching recognition method based on evidential reasoning and interval number.Firstly it determined the similarity evaluation model with multi feature based on D-S evidence theory.Then it generated the basic probability assignment by the interval number method from the historical sample data.And it smoothly converted the weight optimization problem to an optimization problem by the maximum entropy function method.It introduced the hybrid social cognitive optimization algorithm to solve the optimal weights of evidence,and completed the similarity judgment of adjacent fragments by D-S evidence theory.The experimental results show the proposed method is effective.%针对文物碎片特征信息模糊不确定带来的误匹配问题,提出了一种基于证据推理和区间数的相邻文物碎片判定方法.首先确定基于D-S证据理论的多特征相似性评价模型;然后利用历史样本数据使用区间数方法来产生基本概率指派;接着利用极大熵函数法将权重优化问题转换为光滑的最优化问题,采用混合社会认知优化算法来求解最优证据权重,最后利用D-S证据理论实现相邻碎片的相似性判断.实验结果表明了本方法的有效性.