The Study On The New Algoirthms Of Grey Relational Analysis And Its Significances

With the continuous development of science and technology, information has penetratedinto every parts of human life, and become essential elements of modern society. In the era oflife filled with information, people began to realize the importance of how to scientificallycollect, screen, deal with information. Grey system theory is just one of the emergingdisciplines used for information processing. Using mathematical methods, the theory putforward the theories and methods of how to process and analyze systems with incompleteinformation. Because it can describe the behaviors and development trends of things byextracting known information, grey system theory has a wide range of applications. Includingagriculture, military, economic, ecological and other fields. The grey relational analysis is afactor analysis method based on grey system theory. This analysis method can determine therelational degree of the factors which influence the system by considering the similaritiesbetween them. By analyzing and comparing the relational degrees, we can determine thedominant factor which influence the system development. So as to make a quantitativedescription of the system development trend. Compared to other methods of statisticalanalysis methods, grey relational analysis have the following obvious two advantages:(1) thesmall sample size;(2) the ability to study the dynamic process of the system and things. So far,the grey relational analysis has been adopted in various fields to forecasting and analysisingsystem. With the deepening and expansion of the application, a variety of new models andtheories have been further excavation. Based on the existing relational analysis algorithms,and to fit the practical applications, we improved part of the relational models and methods.The experiments show that our work got good results. Its main tasks are as follows:First of all, we briefly summarized the sequence grey relational analysis methods. Wefound that these methods existed rarely can carry on asynchronous comparison. So we putforward the sliding grey relational analysis. The method uses the concept of analogicalreasoning for reference to carry on comparison in asynchronous sequence. Taking JilinProvince for example, we used sliding grey relational grade to study the development trend ofJilin Province’s GDP. This can guide for better development of macro-control policies andprovided new ideas for achieve economic transformation during the "12th Five-Year" periodof Jilin province.Secondly, in practice, things are always described in matrix form. When using greysystem to analyze and process such things, we usually convert the matrix into sequence andthen compare. Such processes would introduce errors and miss many features of things. So weneed to optimize the existing relational analysis methods. In this paper, we deeply analyze ofthe characteristics of grey relational analysis and matrix, put forward to a class of matrix greyrelational analysis method. We successfully use them to provide new technology in thetwo-dimensional signal processing. Such as noise monitoring point position optimization,speech recognition and image processing techniques. And we also proved the rationality ofour methods. From distance "antisense" and grey correlation four axioms, we gave the derivation process of the three matrix grey relational degree we proposed (the absolute matrixgrey relational degree, relative matrix grey relational degree and matrix grey type B absoluterelational degree). Finally the application examples are showed.Finally, due to time constraints, this paper only completed some preliminary research inthe grey relational analysis theories and applications. We did not actually involve the field ofmatrix grey relational analysis. There are also many problems to be explored and dig. Forexample, we only defined sliding grey relational degree and three kinds of matrix greyrelational degree. During these problems, matrix grey relational analysis use in noisefluctuations is still unconsidered; this will be the focus of the future work. In addition, thereare still a lot of prosperities existed in these grey relational degrees, as well as the applicationvalue and character proof. Also how to choose the appropriate reference matrix is urgent needto address the problem in the car of Chinese speech signal recognition and face recognition.