The Research Of Graduate Students’ Comprehensive Quality Evaluation Method Based On AHP And BP Neural Network

The talents play an important role in the development of a country. With the development of our country’s economy, the demand for high-level, high-quality talents is growing rapidly. University graduate students, as high level talented person, take the historical responsibilities to make contributions to our country’s scientific and technological progress, to cultivate innovative talents for the prosperity of our country, to create high level of scientific payoffs, and to provide high level of social services. So, our country should attach importance to improve the quality of graduate education and graduate students’ comprehensive quality level and so on. They have become the top priority in the modern education system reform in our country. University graduate students’ overall quality appraisal is refers to the use of effective methods to collect, sort and analyze the graduate comprehensive qualities, and finally to make a reasonable judgment, ultimately promote the graduate students to improve their comprehensive quality, improve education quality of teaching. This thesis mainly researches around the design of the evaluation system of college graduate comprehensive quality evaluation based on AHP and BP neural network, the main researches is as follows:(1) In order to evaluate the comprehensive quality of college graduate students, on the basis of comparative analysis of other college and university graduate students’ comprehensive quality evaluation index system and evaluation method, carefully analyze the factors that influence graduate students’overall quality appraisal and the influence degree of various factors; eventually establish the index system model of graduate students’ comprehensive quality evaluation, including the ideological and moral quality, professional quality and cultural and sports quality three first level indicators and twelve secondary indicators, and make a detailed analysis of the connotation of the each evaluation index.(2) Using analytic hierarchy process (AHP) to set the universities graduate students’ comprehensive quality evaluation index weights and calculate the graduate students’ comprehensive quality score. First of all, a brief introduction to the analytic hierarchy process, analyze the basic steps of analytic hierarchy process, including the following four steps: Establish a hierarchy, construct the paired comparison matrix, calculate the weight vector then do consistency check, and calculate the combination weight vector then do combination consistency check. On the basis of the analysis, using powerful MATLAB software program, fast and conveniently calculate the relative weight of each evaluation index, and finally calculate the evaluation scores of graduate students’ comprehensive quality, provide sample sets for neural network training and testing below.(3) The realization of university graduate students’ overall quality appraisal method. Firstly, analyze the traditional BP neural network algorithm, and based on this analysis, we advance the corresponding improved algorithm. Secondly, introduce the basic ideas of BP neural network applied to university graduate students’ overall quality appraisal process, and from these aspects, such as:the network layers, the number of neurons in each layer, the neurons conversion function and the determination of learning rate, we design the structure of universities graduate students’ comprehensive quality evaluation model on the basis of BP neural network. At last, by MATLAB simulation, the evaluation results which calculated using analytic hierarchy process are divided into training samples and testing samples, respectively train and test the construction network model, finally gain the more accurate evaluation results in a more convenient evaluation way.To evaluate the comprehensive quality of college graduate students by using AHP and BP neural network, we get the accurate and reliable evaluation result, which opens up a new way for the reasonable evaluation of university graduates’ comprehensive quality, and provides a meaningful reference value.