A Data Mining Method on Software System Layer

This paper puts forward a data mining method on software system layer which is based on researches on clustering data mining. At first, this method gathers all kinds of software data. Then it extracts characters from them to mark softwares on Halstead software science. Thirdly, it classifies the softwares in different categories. For the softwares in the same category, it regards them as the same software cost or the similar software structure. For the softwares in the different category, it finds the essential different factors among them. At last, we show 5414 experiment results in software system data mining. So we can conclude that this method is feasible and effective. Keywordsdata mining; software system layer; clustering