学生评教的数据分析与处理
摘 要
模型通过对所给数据（Access表中）的总体概括分析及个别抽样调查。
针对问题1：总结出评教过程中可能出现的引发不公平判断的因素:
1.少数学生由于个人恩怨产生的对某教师的偏见;
2．不同院系、专业的老师所教的学生数量差别;
3．不同老师授课数目不同.
针对问题2：使用方差分析、F判别法研究了参评人数的多少、教师任课门数的多少对评价结果的影响程度。
针对问题3：通过对数据波动程度因子Si统计研究，分析出了这两个因素对评教稳定性的影响,得出参评人数越多，授课数越少，评分结果越稳定的结果。在定量分析中，应用线性回归模型计算总结出了异常评分剔除率公式y=0.0419+1.0693e-005*x.
针对为体4：对现有的加权平均法的评教模型进行了优缺点的分析，指出需要改进的方面。对各项评教指标进行直观分类，然后通过绘制图表等分析方法检验分类是否合理。之后对每1组指标分析出对应的模糊矩阵和权重向量，得出各组指标的综合评价向量 。再计算整个系统综合评价向量，求出各教师的综合评价分值。然后与加权平均法的结果相互比较，指出新模型的优劣。
关键词:单因素方差分析、波动程度因子、1元线性回归、层次分析法、模糊综合评价
Student comments the data analysis and processing which teaches
Abstract
The model through to gives the data (in the Access table) the overall summary analysis and the inspanidual sample investigation.
Struggle to question one, The factor which summarizes comments teaches which in the process possibly to appear the initiation not fairly judges: one The minority student because inspanidual gratitude and grudges produce to some teachers prejudice, two The different courtyard is the student quantity difference which specialized teacher teaches, three Different teacher teaches the number to be different.
Struggle to question two, the use variance analysis and F distinction law studied the senate to comment population how many, the teacher teaches classes gate number how many appraise the result the influence.
Struggle to question three, Then through the logarithm according to the undulation degree factor Si statistical survey, analyzed these two factors to comment teaches the stable influence, obtained the senate to comment the population many, taught the number to be less, grades a result In the quantitative analysis, the application linearity return model computation summarized exceptionally has graded the rejecting rate formula y=0.0419+1.0693e-005*x.
Struggle to question one, after that is commenting to the existing weighted mean method teaches the model to carry on the good and bad points analysis, pointed out needed to improve aspect. Comments to each item teaches the target to carry on the direct-viewing classification, then through analysis method inspection-classification and so on plan graph is whether reasonable. Analyzes the correspondence afterwards to each group of targets the fuzzy matrix and the weight vector, obtains each group of targets the synthesis appraisal vector . Again calculates the overall system synthesis appraisal vector. extracts various teachers synthesis appraisal minute value. Then mutually compares with the weighted mean method result, points out the new model the fit and unfit quality.
Keywords: Single factor variance analysis; undulation degree factor; a Yuan linear return; level analytic method; fuzzy synthesis appraisal
