Design And Development Of Decision Support System For Students' Evaluation Of Teachers Based On Data Mining

In universities and colleges,the evaluation of teachers' teaching quality is important reference for decision-makers to formulate relevant policies,adjust teaching system and introduce teaching talents.Besides,present approaches to evaluation are mostly adopted based on“questionnaire survey".Only the data on evaluation results are easily and statistically processed,so it is impossible to obtain effective information accurately and objectively.In view of the deficiencies of current teachers' teaching evaluation system such as low degree of automation and unscientific data processing,it is proposed that the data to be evaluated shall be sampled and processed by data warehousing,online analytical processing and data mining technologies.A data warehouse and a theme-oriented multidimensional database are constructed pursuant to the data from student performance management system and teachers' teaching evaluation database managed by the Office of Education Administration.Besides,centering on a fact table,a source database for online analytical processing is established by a star model from the perspectives of students,teachers,courses and time.An emphasis is laid on discussing CART(classification and regression trees)algorithm centered decision tree generation algorithm and pointing out the disadvantages of the algorithm.Then,this algorithm is improved according to the correlations among its attributes.Besides,it is proved by a test that improved algorithm may effectively shorten processing time and reduce the size and amount of generated rule set.Meanwhile,knowledge is mined from teaching evaluation data by the improved algorithm and decision rules are obtained to assist the school in improving and making related decisions.Based on the procedure to construct a data warehouse and online analytical processing process,a multidimensional processing system based on data mining is designed and developed to evaluate teaching in combination with the decision tree algorithm for mining data on the basis of browser/server mode.To develop this system,a framework is designed as well,and a theme-oriented data warehouse is established by extracting relevant data from educational management database,and databases of student management system and teacher management system.Then,the data in databases are processed by online analytical processing and mined.By data mining technology,impacts of different factors on teaching evaluation can be discerned to better help with school administrators in making related decisions.