Research On Classification Of Colleges In Our Country Based On Clustering Technology Of Data Mining

With the development of higher education of our country, the amount of colleges is large, the type of colleges is complex, if can't classify management and development of colleges, will restrict the development of higher education of our country. The current classification of colleges in China is leading by government, has a strong policy and subjectivity; the classification of domestic academic institutions is unrealistic; and the classification of foreign academic institutions can't be fully applicable to the condition of our country. With the successful application of data mining technology in the financial, telecommunications and other areas, this research uses clustering technique of data mining for classification of colleges in education.First, this thesis proposes the classification criteria from both teaching and research from the angle of college's function, and chooses the appropriate elements of the basic characteristics of college's classification determine the weight of each indicator, build a KPI system relatively complete and reasonable, from "the database system of the basic teaching statement of the national colleges". Secondly, this thesis preprocessing the raw data of "the database system of the basic teaching statement of the national colleges", including data source access, data cleaning, data integration, data transformation, data reduction, etc, to make it more consistent with the standards and norms of clustering algorithm. Then this thesis uses K-means algorithm and its improved EM algorithm, uses SQL Server 2005 Business Intelligence Development Studio, establishes the model of clustering. According to the dataset, gets different clusters through clustering analysis, to obtain the distribution of college object. Finally, this thesis summarizes the results of clustering, analyzes its feasibility and rationality, obtains preliminary conclusions of this thesis, and analyzes the factors that affect the results and the improvement measures. Furthermore, it analyzes the application value of the clustering results from the university itself, education authorities and the public, also makes some constructive advices in the development of higher education.