基于小波过滤方法的新k-部排序本体算法（英文）

A New k-Partite Ranking Ontology Algorithm via Wavelet Filtering Technology

目的将本体结构图划分成k个部分,利用k-部排序学习得到一个得分函数,从而两本体概念之间的相似度可通过它们之间得分的差值来计算。方法研究AUC标准下基于k-部排序的本体算法。将小波过滤技术融入到本体迭代算法,通过小波的N项逼近来控制顶点的划分。结果将算法应用于基因本体和物理教育本体,利用P@N对结果进行评价并与以往算法得到的结果进行对比。发现随着N的增大,算法的准确率明显高于其他算法。结论实验结果说明新算法对于本体相似度计算和本体映射的建立是有效的。

Objective The ontology vertices are usually divided into kparts,and then the score function is obtained by k-partite ranking algorithm,so the similarity between two concepts is determined by the difference of their scores.Methods This paper tend to study the k-partite ranking based ontology algorithm under AUC criterion.Combined the wavelet filtering technology with the ontology iterative algorithm,and the vertex partition is controlled by Nterm approximant.Results New algorithms are applied in gene ontology and physics education ontology,and we use P@N to measure the result data and compare them to the results from former algorithms.It is implied that the accuracy of our algorithm is significantly higher than that of other algorithms as the increase of N.Conclusion The experimental results show that new algorithms are effective for ontology similarity computation and ontology mapping.