【摘要】：在經(jīng)典的變異評(píng)分計(jì)算過程中,因?yàn)椴豢紤]被播種軟件缺陷的檢測(cè)難度而使得變異評(píng)分的可信性受到質(zhì)疑.因此提出一種基于缺陷檢測(cè)難度評(píng)價(jià)測(cè)試用例集合的方法.以logistic回歸為基礎(chǔ),利用經(jīng)驗(yàn)回歸方程建立缺陷的識(shí)別概率與缺陷檢測(cè)難度之間的數(shù)量關(guān)系.借助關(guān)系曲線下的面積,變異評(píng)分被重新定義.新定義的變異評(píng)分不但不受缺陷樣本的檢測(cè)難度影響,而且規(guī)避了因等價(jià)變異體的出現(xiàn)而使得經(jīng)典變異評(píng)分的計(jì)算不準(zhǔn)確的問題.
[Abstract]:In the classical variation score calculation process, the credibility of the variation score is questioned because it does not consider the difficulty of detecting the defects of the sown software. Therefore, a method of evaluating test case set based on defect detection difficulty is proposed. Based on logistic regression, the quantitative relationship between defect identification probability and defect detection difficulty is established by using empirical regression equation. The variation score is redefined with the aid of the area under the relational curve. The newly defined variation score is not only not affected by the difficulty of detecting defective samples, but also avoids the problem that the calculation of classical variation score is inaccurate because of the appearance of equivalent variants.
【作者單位】： 北京郵電大學(xué)網(wǎng)絡(luò)與交換國(guó)家重點(diǎn)實(shí)驗(yàn)室;北京郵電大學(xué)自動(dòng)化學(xué)院;
【基金】：國(guó)家自然科學(xué)基金項(xiàng)目(61202080) 廣西云計(jì)算與大數(shù)據(jù)協(xié)同創(chuàng)新中心 廣西高校云計(jì)算與復(fù)雜系統(tǒng)重點(diǎn)實(shí)驗(yàn)室資助(YD16508)
【分類號(hào)】：TP311.53
，

本文編號(hào)：2222674