Research Of Flexible Integration Technology And Application In Hydrological Data Mining

The data analysis system has the characteristic that the core function unit changesdynamically because the user demand is changeable. The development, applicationand maintenance of these systems are difficult. The repetition of development iscommon. Therefore, the integration of analysis system needs to research a frameworkwhich is flexible in certain to serve for it urgently.This paper takes analysis systems such as the data mining as the research objects.The flexible integration theory and the technology are researched and analyzed,including the EAI integration pattern, the component technology, the wrappertechnology, etc. Then according the flexible integration requirements of the analysissystems, a flexible integration concept fits to develop and apply these systems is given,and a flexible integration framework is designed. Besides these, this paper carried ona detailed analysis and elaboration of function mechanism in each level of theframework..What we do in this paper goes as:1．Discussed the flexible integration demands and question existed in the currentanalysis system, and researched the flexible integration related theory and technology,analyzed several kinds of flexible definitions, as well as to related flexible integrationtechnologies including EAI model, component technology and wrapper technology.Then based on these analyses, we gave the flexible integration concept of analysissystem.2．Under the given flexible integration concept, flexible integration mechanismwas disscussed, a wrapper which can wrap the isomerism modules and an integrationframework which customized the process management according to the user demandwere designed. The focus of this paper is the research on wrapper model andcustomization flow management mechanism; some other function modules of theframework are briefly analyzed.3．Based on the flexible integration framework and combined with the hydrologydata mining domain characteristic, we finished a hydrology time series data miningflexibility integration experiment, and the confirmation of the usability of the flexibleintegration mechanism and the framework was also did.