Research On The Evaluation Of Undergraduates’ Theoretical Foundation And Application Skills Based On FLFM Clustering Algorithm

Due to the impact of the social environment, educational philosophy, teaching methodsand other factors in China, resulting in practical teaching of the university have become agreat challenge facing our colleges and universities. China’s current teaching model, theteacher pay more attention to explain the theory, but obviously overlook the cultivation ofstudents’ practical ability. Even some teachers want to focus on students’ practical ability, butbecause of the lack of teaching resources, teacher experience and other factors limited,teachers’ idea cannot come true. Graduates’ theoretical basis and application capabilities havea bigger difference. First of all, there should be a scientific evaluation methods, based onwhich can the educational reform be confirmed and conducted.The existing assessment method is relatively simple for college graduates’ evaluation indomestic universities, it mainly calculates graduates’ average score of each professionalbackbone course, and then it is concluded that the average score from high to low sequence,evaluation results are obtained. It is clear that this evaluation method could not truly reflectthe specific differences between basic and applied theoretical capacity of students.Traditional clustering algorithms such as k-means algorithm needs to give clusteringcenter and clustering number at first, Graduates have a large number of courses taken, thescore are widely distributed, it is a high-dimensional problems, so it is very difficult to clearthe clustering center and the number of clusters, choosing a different cluster centers will makedifferent clustering results. In addition, determining the number of clusters is not an easything, mainly determining the number of clusters is not strictly theoretical basis, it isunrealistic to determine the number of clusters by relying on people’s subjective sense. FCMalgorithm also exist similar problem. SOFM algorithm which the clustering center does notneed to give in advance, but it is obviously insufficient that network training needs a longtime, some nodes can’t always win in the competitive layer, which can cause the clusteringresult is not accurate. This several clustering algorithms above do not apply to the evaluationquestion of the college graduates’ theoretical basis and application ability. Aim at aboveproblems, this paper applies the Fuzzy Logical Feature Fapping Artificial Neuron Network tothe evaluation of the graduates’ theoretical basis and application ability, The mining andanalysis of students’ achievements of the course and by other clustering algorithms (k-meansalgorithm, FCM algorithm and SOFM algorithm) were compared to verify the effectivenessof the algorithm on this issue. It has many features, for example, neural network trainingspeed is faster, the clustering center and the number of clusters do not need to give in advanceand robustness is relatively strong (anti-noise ability is strong) and so on. At present the algorithm is needed cluster analysis for a grade of school of computer, it iseasy to spread to the scope of the school, it can analysis and process integrated and largeamounts of data for science, arts, normal, non-normal and all the students, and gives acomprehensive, objective, and scientific evaluation on the graduate students’ abilities, and toprovide a significant and precise decision basis for the university decision makers.