Rough Set Theory In Machine Learning Applications And Research

Learning is a unique human ability,it is a experience of accumulation in order to improve its performance in the process.Machine learning is the second expert after another important application of artificial intelligence research, also it is the core of the research of Artificial intelligence and neural computing. The basic goal of the Machine learning is to have the computer ability to learn, simulation or the realization of human learning, it is the study of computer recognition of existing knowledge, acquiring knowledge and seeking to improve its performance and achieve their ideal solution.In many areas machine learning has a successful use and machine learning has become a new discipline on the brink. Integration of a variety of learning and a variety of integrated learning system is rising. A variety view of machine learning and artificial intelligence-based is taking shape; a variety of learning methods is expanding in application, part of the application of research findings have been turned into commercial products; Machine learning and academic-related activities more active than ever before.Rough set theory is a new mathematical tool which does not have the integrity and the uncertainty for information. The main idea is to ensure the classification ability of Knowledge and to reach the decision-making rules or classification of the problem. The theory has been widely used in data mining, machine learning, process control, analysis and decision-making in areas such as pattern recognition.The decision tree have a very widly use in the range of applications in machine Learning. Decision Tree is an important method of inductive inference learning. This paper proposes a decision tree algorithm based on rough set. Firstly we use the classification accuracy and the certainty factor of the decision-making rules in the structure of the decision tree.There are inhibition factors in the process of forming Algorithm, cuting branches for decision tree, avoiding redundant steps of cuting branches. Secondly we will match the conditions of property value and decision-making property values in each division to avoid unnecessary calculation and improve the speed of the algorithm.On the other hand, to the database of dynamic changes, based on the analysis of classic method. Rough set theory will be applied to obtain the rules of the system, to obtain incremental learning based on a rough set of rules, made incremental learning process,when the new rules joined the targets of Knowledge. The algorithm make full use of the existing rules in the Knowledge Base, to reduce the scope of the search space, so as to make a method of decision-making value updating,and give full play to its emphasis on decision-makers Analysts and the importance of the initiative to meet the ever-changing external circumstances, rather than rely solely on the decision-making machinery in order to demonstrate people's decision-making and machine learning the interaction between.