Research On Evaluation Model Of University Teacher’s Research Ability Based On Rough Set And SVM Technology

With the development of society and technology, more and more teachersparticipate in the school’s research on the basis of teaching, with the research and practiceto promote teaching. Now, teachers are becoming one of the major forces in researchactivities, and the strength of the activities has also became a symbol to measure theoverall strength of school. Therefore, how to evaluate teachers’ scientific research abilitycorrectly has become a major problem in current colleges and universities.The main work as follows:Firstly, overview the current related research on the evaluation model of teachers’research ability which were established by scholars at home and abroad. Then describethe main technical methods which used in the model and summary the defects, based onthis to submit the main technical methods in the model of this paper.Secondly, mainly discuss the theoretical methods used in the evaluation model, suchas the neighborhood rough set theory and BP neural network theory and algorithm,support vector machine theory and particle swarm optimization algorithm and so on.Thirdly, pretreatment the initial sample data which is collected from thecorresponding scientific data on the basis of the established evaluation index system, andthe processed data input for BP network and SVM algorithm to set up the evaluationmodel of teacher research ability. Explain the merits of BP algorithm and SVM algorithmused in the evaluation model in term of the prediction accuracy and prediction error, anddescribe the reason why we select SVM algorithm to build teachers’ research capacityevaluation model.Fourthly, regarding to the multi-index evaluation system, we propose to simplifythe evaluation index through the neighborhood rough set attribute reduction, lower thedimension of the evaluation, to improve the efficiency of vector machine algorithm andspeed up the model’s building; for the selection of SVM parameter, we propose tocombine the particle swarm optimization algorithm with support vector machinealgorithm to further optimize the evaluation model and improve the accuracy of itspredictions.Fifthly, Summarize the development background, design architecture, environmentof system development and related technology of the scientific research managementinformation platform, develop and achieve evaluating function of the model, show some results of research evaluations in the system.With the development of information technology in colleges and universities, theteachers’ scientific research ability evaluation is more and more important, in the paperpropose a new method based on rough set theory and particle swarm optimizationsupport vector machine technology which improves the training speed and predictionaccuracy of the model and provide scientific basis for the application of evaluationmodel.