A Study On Decision Tree Algorithm And Its Application In Collection Management Of Urban Construction Archives

As a trend of the urban construction archives develops in the future, digital urban construction archives will surely accumulate much more digital archives information. Because data mining techniques develops rapidly, more and more researchers combine data mining techniques with urban construction archives management to discuss and research. In order to improve the blind and passive problem of current collection work of urban construction archives, we present in this paper an improved C4.5 algorithm which uses the RainForest framework, apply this algorithm to archives collection work of the urban construction archives for scientific guidance and gain satisfying effect. The main content is as follows:Firstly, research various classic decision tree algorithms, analyze and compare the advantages and disadvantages of the different decision tree algorithms and finally select the C4.5 algorithm which can process continuous-valued data and has simple tree structure.Secondly, apply RainForest framework and effectively improve the disadvantage that traditional C4.5 algorithm has poor scalability. Analyze and compare the memory occupation between traditional C4.5 algorithm and improved C4.5 algorithm in aspects of weather data example and prove in theory.Finally, apply improved C4.5 algorithm to collection management of urban construction archives, design the urban construction archives management information system in software engineering's development flow and object-oriented thinking, and realize it with Hibernate, SWT/JFace and other technologies.