一种提高检测效率的V-detector优化算法的设计

Design of optimized V-detector algorithm with higher detection efficiency

实值阴性选择算法V-detector在产生检测器时不能确保检测器具有较大的覆盖范围,其结果是检测器集合中的检测器数量过多,检测效率较低。为提高检测效率,提出了V-detector优化算法,一方面,通过合理确定检测器的中心点位置及检测半径,扩展了检测器的覆盖范围;另一方面,优化算法采用假设检验的判定方法判断检测器集合对非自体空间的覆盖率。假设检验融合到检测器集合的生成进程中,在确保检测器集合满足覆盖率要求的条件下,减少了检测器集合中的检测器数量。实验结果表明,与原算法相比,优化算法使检测器集合中检测器的数量大幅度下降,检测效率得到提高。

V-detector algorithm is unable to guarantee a large coverage for generated detector.The shortcoming usually results in a large number of detectors in the detector sets and low detection efficiency.An optimized V-detector algorithm was proposed.In the optimized algorithm,a detector was generated on the basis of a randomly chosen uncovered non-self sample,and the detector generation process ensured the detector covered the non-self sample and extended the detector＇s coverage by determining the detector＇s center and detection radius reasonably.Meanwhile,hypothesis testing process was used to estimate the coverage of a detector set,and hypothesis testing process was integrated with detector generation process to ensure the generated detector set fully covered the non-self region,and reduced the number of detectors in the detector set.Experimental results indicate that compared with the V-detector algorithm,the optimized algorithm generates a much smaller detector set and increases the detection efficiency.