The New Methods Of Data Mining And Its Application In Complex Industrial Process

Data analysis is the basis tasks in data mining. It also is the basic of scientific research in data mining, many data mining methods are based on the different methods of data collection and analysis. Relevance vector machine is a novel kind of machine learning method in recent years, it has many advantages, but has many unresolved issues too, the principles and the algorithm is discussed in detail. In order to optimize the control of complex industrial, the complex control process based on data mining is studied, the new control strategy with complex industrial is given. Rough set theory in data mining is an important theoretical knowledge, and attribute reduction in rough set theory is the core problem, various new attribute reductions is proposed, many novel rough set attribute reduction ideas and methods is developed, and many experiments is studied. The new methods and ideas are researched in data mining, these worked for the promotion of the data mining methods.The main work is as follows:1. A new approach was proposed, called as discrete differential evolution algorithm for solving attributes reduction. The differential evolution algorithm is researched for solving attributes reduction in rough set. And, the improve strategy for the proposed, the contrast of performance is prepared.2. A new though for solving attribute reduction based on Discrete Particle Swarm Optimization Algorithm in the rough sets was proposed. The Discrete Particle Swarm Optimization Algorithm is researched for solving attributes reduction in rough set. And, the improve strategy for the proposed, the contrast of performance is prepared.3. A new attribute reduction algorithm based on pruning rules was developed. The pruning thought is studied in Functional Dependency, the though of pruning rules and pruning ideas in the attribute reduction of rough sets was proposed, furthermore, many examples of reduction was given. The experimental results demonstrate that the developed algorithm have good property.4. A new algorithm of rough set attribute reduction based on PSO was proposed. Combining the theory of rough sets reduction and the algorithm of particle swarm optimization (PSO). The attribute reduction algorithm based PSO was developed.5. A new algorithm of rough set attribute Reduction based on Bisection Method was proposed. Although, the algorithm is high complexity, it is complete. And the Cache thought in compute is introduced into the algorithm.6. The Relevance vector machine is studied, and the new data mining method is introduced into the complex industrial process. The Features and using methods on the complex industrial process is discoursed, the optimization strategies with complex industrial process is researched.7. The prospect and the research aspect of data mining is discussed, and the improvement and shortage of many attributed reduction is presented, these are the good tries in data mining.