Bank Loan Classification Preliminary Study Based On The Support Vector Machines

This paper will use support vector machine to classthe bank load. The tradition method of classification are four classification methods,which merge the weight and the financial indexes,and finally the number less than 100 will be obtained in order to class the load. Because our paper makes use of the new five classification metheds of People's Bank of China, and its weights and parameters should be used in manufacturing, and the financial indexes method may transform the multiple dimension data to one dimension data,the computing procedure is very complicated. So the support vector machine is applied to class the load,the expansibility and veracity are also excellent.Our paper introduces the theory of support vector machine, in order to master the concept and method in support vector machine, which will be helpful in practice. In application part of bank load classification, we do not want to discass this part in details,because it is very small part in support vector machine application.Machine learning is the study of how computer simulation or the realization of human learning works, in order to obtain new knowledge or skills to reorganize the structure of existing knowledge so that they continue to improve its performance. It is the core of artificial intelligence, and it has a smart computer fundamental way, its application is across all areas of artificial intelligence, mainly into the use of summing up, rather than a comprehensive interpretation. Machine learning research is based on the physiological, cognitive science, and so on the understanding of the mechanism of human learning, the establishment of the learning process of human calculation models or model of understanding to develop a variety of learning theory and learning methods, general learning theory, algorithms and analysis. The establishment of task-oriented is with a specific application of the learning system. The purpose of machine learning is based on the given training samples of input-output system dependent relationship between the estimated output of the unknown that it can make the forecast as accurate as possible. Dependent relationship between input-output is called the study function, learning model of machine learning. Pattern Recognition (Pattern Recognition) is the characterization of the things or phenomena in all its forms (values, language and logic of relations) for information processing and analysis of things or to describe the phenomenon, identify, classify and explain the process . Information science and artificial intelligence is an important part.This paper discussed is the use of computer pattern recognition which is the realization of human pattern recognition ability. Its theories and methods are more universal. The role of pattern recognition is to allow a computer in the face of things, when its right is under a certain category. Pattern recognition research is focused on two aspects, one (including people) is how to percept object which belongs to the scientific understanding of the scope, and the other is how to use computer pattern recognition theory and methods in given task. The purpose of pattern recognition is the classification of unknown samples with attribution. By means of chemometrics methods, pattern recognition is able to achieve a direct measurement of the physical nature of the implicit recognition. With the computer in the 1940s, as well as the emergence of the 50's the rise of artificial intelligence, of course, it is hoped that people can use computers to replace or expand some of the human brain. Computer pattern recognition is developed rapidly in the early 1960s and becomes a new subject.With other things, the emergence of computer improves the development of pattern recognition for practical applications, which, in turn, it further develops its theory. Industrial production automation and information processing make the pattern recognition develop in advanced stage of today's engineering and research applications. For decades, Pattern Recognition has achieved a great deal of research results in many ways, and the success of the application promotes the development of artificial intelligence. Pattern recognition and statistics, psychology, linguistics, computer science, biology, cybernetics, and so have the relations. Pattern recognition and artificial intelligence, image processing are researched in overlap mode. For example, the adaptive self-organization of pattern recognition system includes a mechanism to study artificial intelligence; artificial intelligence research of understanding the landscape and natural language understanding also includes pattern recognition problem. Another example is pattern recognition in the pre-feature extraction and image processing applications link in the technology; image processing image analysis inclides application of pattern recognition technology also.Of course, the application requires models and data authenticity, due to capacity constraints, data extraction would be comprehensive and dimension have reached 12, many factors make it likely not be an ideal model. But perhaps the surprising results will be ideal, this is the realization of the unknown before. Perhaps there is in the process of the emergence of a new, short, the hope for support vector machines in our modest contribution to the development.At present, there are many methods to classify the bank loan, and has the good and bad points respectively. In this paper we use support vector machines to carry on the classification to the bank loan, introduced the pattern recognition application elementary knowledge briefly, made the system, the entire introduction to support vector machines, enable us to have a more thorough understanding to it, grasps its work mechanism, lay the foundation for this article core work.Has given the exhaustive explanation to Support Vector machines nuclear function , made the good preparation for the following experiment process. Also has carried on system's introduction to bank loan's 5 classification class method. This article has made the preliminary research work based on the Support Vector machines bank loan classification, And has carried on the experiment using different nuclear function to the research content with the C language write program, obtains the experimental result data, and analysis the experimental result, this article has obtained the more ideal experimental result based on the Support Vector machines bank loan classification,and effectively. And we has made the summary and the forecast bank loan classification based on the support vector machines,will be advantageous for later refers in this aspect other researchers.