Research On Property Mining Methods Based On Genetic Algorithm

This paper studies the principles and actuality of data mining; discusses the corresponding concepts, working steps and key technologies on data mining; analyzes and investigates thoroughly the essential theory of rough set and genetic algorithm, basic methods and algorithms of attribute reduction. By studying deeply the common realization problems of genetic algorithm, the design and realization method for various operators is proposed;proposes improving of the code,during coding ,not use consistent code,but use various code,Then the individuals in population are selected by roulette wheel algorithm,improves the possibility of the methods at large; Improving end condition to be determined by population stability can assure hidden parallelism of genetic algorithm, which can make the algorithm not only carry on the highest reduction for decision information system, but also achieve reductions as more as possible. The test proves that the algorithm can achieve all the reduction for most of testing data sets, process relative reduction problems efficiently and calculate reduction for inconsistent information system.