期刊名称：International Journal on Computer Science and Engineering
印刷版ISSN：2229-5631
电子版ISSN：0975-3397
出版年度：2010
卷号：2
期号：5
页码：1754-1757
出版社：Engg Journals Publications
摘要：Billions of lines of code are currently running in Legacy systems, mainly running machine critical systems. Large organizations and as well as small organizations extensively rely on IT infrastructure as the backbone. The dependability on legacy Software systems to meet current demanding requirements is a major challenge to any IT profession. One of the top priority of any IT manager is to maintain the existing legacy system and optimize modules where required. Various techniques have been developed to determine the complexity of the modules as well as protocols have developed to assess the severity of a software problem. In this paper, it is proposed to study data mining algorithms in a multiclass scenario based on the severity of the error in the module.
关键词：Legacy software; Normalization; Data mining; Random tree; Bayesian Logistic Regression; CART.
