CAJ | 학술논문

设计了一种应用机器学习和D S证据理论来进行 Linux病毒检测的方案。主要包括方案的总体框架、样本特征选择方法、分类器选择、检测效果融合以及方案验证与结果分析等。在样本特征选择时引入了控制流程图的概念，在检测效果融合时使用了D S证据理论的方法。最后在基于 Weka软件的机器学习平台上实现和测试了该方案。验证结果表明，该 Linux病毒检测方案具有良好的检测率和可靠性，可以应用于实际的商业产品中。
설계료일충응용궤기학습화D S증거이론래진행 Linux병독검측적방안。주요포괄방안적총체광가、양본특정선택방법、분류기선택、검측효과융합이급방안험증여결과분석등。재양본특정선택시인입료공제류정도적개념，재검측효과융합시사용료D S증거이론적방법。최후재기우 Weka연건적궤기학습평태상실현화측시료해방안。험증결과표명，해 Linux병독검측방안구유량호적검측솔화가고성，가이응용우실제적상업산품중。
This paper mainly designs and realizes a Linux virus detection method using machine learning and D-S theory.It includes the design’s general framework,feature selection method,classifier selection method,detection result fusion and the design verification and result analysis.It intrdouces the control flow graph while doing feature selection,and introduces D-S theory while doing detection result fusion.Then it implements and test the method on the platform of Weka software.The results of implementation show that this design to detect Linux virus has high efficiency and good reliability,and it is adequate for commercial products.