Canonical Correlation Analysis In Data Mining Application

This paper focus mainly on the application of the multivariate statistical methods in Data Mining。The method is based on the canonical correlation analysis of multivariate statistical method。This research mainly adopts the theory to combine with reality, innovating on the basis of traditional research methods, Combing with the example analysis by using computer to realize research methods。The main research contents is mathematical statistic theory, methods, practice, data mining theory, technology, arithmetic and computer software。The detailed implement scheme:firstly, learn theory and computer software systematically; Secondly, read the relevant documents; Thirdly, confirm the framework and theoretical part of the paper; Finally, combine with the computer to solve the problems by analyzing the actual problems, choosing the appropriate method。The main methods of multivariate statistical analysis methods include:the principal component analysis,discriminant analysis,cluster analysis and factor analysis, canonical correlation analysis, etc。This paper introduces the content of the basic idea of canonical correlation analysis,the steps of canonical correlation analysis,the nature of canonical correlation variables,the significance tests of canonical correlation coefficient and so on。It also introduces the generation, definition,process,application and development of data mining。Data mining and mathematical statistics are interpenetrative, so data mining's development is inseparable from the mathematical statistics。Similarly, data mining not only poses a challenge to the development of mathematical statistics but also provides the opportunity。Finally, some useful conclusions have been drawn through systematic analysis of the students'grades which they have got during the four years'study in Changchun Industrial University。They are from basic science college information and computing science department, grade 2006。Based on the analysis of student achievement for data, applying the canonical correlation analysis in statistics into data mining。Dig out the rules with certain credibility, which is of important significance。These conclusions can be helpful to the future development of students' learning。At the same time the teachers can also take extra-curricular studies of the course of some relevant knowledge properly, according to the relevancy of their courses and other courses, which provides some help to the teachers' teaching。...