The Application Of The Teaching Evaluation System In The Data Mining Technology

The effective evaluation of teaching quality can make the teaching managementdepartments grasp the teachers’ status and the conditions accurately, while it also providesbasis for teachers to improve teaching work, develops teaching reform and teachingmanagement department to improve the management conditions. So how to use the effectivetechnical in the scientific, comprehensive and fair way, which is very important no matter onthe theory or in the practice.This paper designs a prediction system based on Association Rule Algorithm andMultiple Linear Regression Model. The establishment of this system will further enhance thecredibility and efficiency in teaching quality evaluation and prediction. The system adopts3layers architecture combines C/S with B/S. It contains four modules: basic data processing,user management, data processing, and the prediction.The module of data processing uses an improved Apriori algorithm to mine frequentitems to construct strong association rules. Based on the strong association rules, the teachingevaluation index system is Optimized and the teaching evaluation data is cleaned. Theimproved Apriori algorithm generate a new association rule during the connecting operation,it will produce a large number of repeat or a strong association rules, each statistical supportneed to be repeated access to the database, the improved Apriori algorithm by addingconditions, prevent from repeating or a strong association rules, so it is reducing the waste oftime and space. The author increases a record in correlation structure in order to support theaffairs index, As forwhether is strong correlation, it is just need to record the number ofelements, so that reduce the access for I/O and improve the computational speed.The main job of forecasting module uses the data after cleaning to establish teachingevaluation prediction model based on regression analysis algorithm, while do a test andprediction. In this paper, there is a comprehensive improvement to select the independentvariable. That is making independent variables partial F check value and using sequeningmethod to its ascending order. Then in half position according to F test do forward cuts or backward increase, so to reduce the independent variable choice in the linear regression. Afterthe selection of independent variable, the optimization model can do a better performance byexperiment.