Research On Classifing Algorithm Base On Rough Set And Genetic Algorithms

Data Mining (DM) is a technique that aims to analyze and understand large source data and reveal useful knowledge hidden in the data. From the different task of DM, there are several patterns as follows: Classification Pattern, Prediction Pattern, Association Rule Pattern, Regression Pattern, Clustering Pattern, Time Series Pattern, and so on. Algorithm is very important for any pattern.The emphasis of this thesis is the algorithm of Classification Pattern. The main contribution of this dissertation includes:1) Survey the current algorithms of classification pattern.2) Introduce the basic theory and concept of rough sets and genetic algorithms, and summarize the classic algorithms of attribute reduction based on rough sets.3) Introduce a dependability algorithm of Attribute Reduction, and then improve the algorithm based on dependability and importance.4) Based on the research of genetic algorithm and rough sets, propose a classification algorithm, and use some data to verify it. Among the algorithm, improve the simple genetic algorithm by changing the probability of crossover and mutation in dynamic.The thesis is organized as follows: Chapter 1 introduces some basic concepts and the development state of Data Mining. Chapter 2 introduces some basic concepts of Rough Sets and gives the survey on several typical reduction algorithms. Chapter 3 introduces the basic concepts of genetic algorithm. Chapter 4 introduces the improvement of the dependability and importance algorithm of Attribute Reduction, and then describes the idea of algorithm based on Rough Sets and Genetic Algorithm. Chapter 5 uses some data to verify the availability of the algorithm.