Research On Neural Network And Its Application In Students' Comprehensive Quality Evaluation

Nowadays,as the talents competition is the key competition of the world,it is important to cultivate the high quality comprehensive talented students,and the students'comprehensive quality evaluation can exhaustively evaluate students'comprehensive quality and improve the development of the students.Our country has long-term influenced by the negative impact of the "examination oriented education",so the students' evaluation is mainly in order to test the learning results,so that to focus on students' mastery of knowledge and skills,the evaluation method strengthens the investigation of the intelligence and the improvement of the examination skills,rather than promoting the students' comprehensive quality improvement.Contemporary society needs the talents with high quality and strong ability,promoting the comprehensive education has caused the government's attention and done some reform,but it also has some problems in the practice,such as focus on results than process,evaluation work concentrated together,it lead the comprehensive process to an empty form and cannot reach the purpose of actualized students' comprehensive quality evaluation.Therefore,it is of great significant to carry out comprehensive quality evaluation and promote students' comprehensive and personality development.Based on the studies of the current situation of the development of the students'comprehensive quality evaluation,the thesis analysis the shortcomings of the current evaluation methods,and establishes the evaluation system of students'comprehensive quality evaluation,and proposes a method that using genetic algorithm converged BP neural network hybrid algorithm for student' comprehensive quality evaluation.And then apply it to the students' comprehensive quality evaluation system.The followings are the main work of the thesis:(1)The thesis stands on the starting point of the real problems existing in the quality of education,combined with related literature,analysis the exits students'comprehensive quality evaluation methods.And through theoretical analysis and practical investigation,the thesis determines students' comprehensive quality evaluation index and establishes the evaluation system.(2)Point at the demand of students' comprehensive quality evaluation,the thesis designs the structure of BP network.And in order to solve the shortcomings such as slowly training speed and inclined to immerse into the local minimal point of the BP network,there is another way that combined genetic algorithm with BP network which is called hybrid algorithm.By using the genetic algorithm global searching capability to optimize the weight and threshold,the network can have a better initial value,to ameliorate the shortcomings of BP network.(3)Based on the functional requirements of students' comprehensive quality evaluation system,designed and implemented the students' comprehensive quality evaluation system,recorded the representative performance into the files to reflect the development evaluation thought,and also implemented the application of BP network in the students' comprehensive quality evaluation system.