基于本体的综合加权案例相似度算法研究

Ontology-based comprehensive weighted similarity algorithm research

针对现有相似度算法存在查全率和查准率不高的问题,提出了一种基于本体的综合加权案例相似度算法。综合考虑案例的各个属性,分别采用不同的属性相似度计算方法,根据各个属性的重要程度加权,计算出案例的相似度。以三对不同类型的心理咨询案例为例对综合加权相似度算法进行了验证,并与传统的基于最近距离的相似度计算方法进行对比,结果表明综合加权相似度计算方法更加精确。

In view of the low accuracy and recall rates of the existing similarity algorithm, this paper presented an ontology- based comprehensive weighted similarity algorithm for case retrieval. Considering the various attributes of the case, it used dif- ferent attribute similarity calculation methods, and weighted according to the degree of each attribute, then calculated the similarity of the case. It validated the similarity algorithm by three psychological counseling case. Finally, by comparing the traditional similarity algorithm based on the nearest distance and the weighted similarity algorithm, it drew a conclusion that the weighted similarity algorithm is more accurate.