摘要
针对历史评价数据时效性会影响评价计算准确性的问题,在前期研究的基础上,对原有方法进行了扩展,引入了评价数据的时间属性,构造了评价数据衰减因子。衰减因子并不直接作用于评价数据,而是改变了评价数据的统计权重,减小了时效性对于评价计算准确性的影响。实验结果表明,评价数据的时效性严重影响评价计算的准确性,扩展的算法在减小时效性影响方面是有效的,算法动态适应能力得到了提升。算法应用于学生评教数据的处理,取得了良好的效果。
According to the problem of freshness for historical evaluation date effection evaluation computation,based on the past research,foregoing method is extended.Time attribute of evaluation data is introduced to construct reduction-factor.Reduction-factor that modifies the statistical weight but not the data itself and reduce the effect of evaluation data freshness is introduced.The experiment proves that the freshness of valuation data impacts the evaluation computation and the new algorithm is effective to treat the effect of evaluation data freshness,dynamic adaptability is improved.This algorithm is introduced into teacher evaluation system to treat the original data,the effect is satisfactory.
出处
《计算机工程与设计》
CSCD
北大核心
2010年第3期684-686,F0003,共4页
Computer Engineering and Design
基金
上海市教育委员会科研创新基金项目(10YZ193)
上海市市本级财政部门预算基金项目(1138IA0005)
上海高校选拔培养优秀青年教师科研专项基金项目(slx08017)
关键词
评价数据
评价计算
时效性
衰减因子
实验分析
evaluation data
evaluation computation
freshness
reduction-factor
experiment
