一种优化深度神经网络的高校教学质量预测

An optimized deep neural network for university teaching quality prediction

高校青年教师的教学水平是影响教学质量提升的关键因素,也是当前高校教学评估的重要内容之一。为了更加准确地评估高校青年教师教学质量,文中设计一种基于遗传算法优化的深度神经网络评估预测方法。该方法首先对现有文献资料进行分析,结合高校青年教师教学实际情况,确定高校教学质量评估影响因素;然后按照影响因素对高校青年教师教学质量数据进行采集分类,从而获取学习样本;最后建立网络预测模型,在教学质量评估一级指标与二级指标之间构建子网络模型,并在子网络输出与教学质量预测输出之间构建深度神经网络,通过遗传优化算法调整网络的超参数,使得网络预测性能显著提升。仿真结果表明,与已有网络模型相比,所提出的深度神经网络评估预测方法在高校青年教师教学质量预测上的准确度更高,具有广泛的教学管理应用价值。

The teaching level of young teachers in universities is the key factor affecting the improvement of teaching quality,and it is also one of the important contents of current teaching evaluation in universities. A deep neural network evaluation and prediction method based on genetic algorithm optimization is designed to evaluate the teaching quality of young teachers more accurately. The existing literatures are analyzed,and the influencing factors of teaching quality evaluation in universities are determined based on the actual teaching situation of young teachers in universities. The teaching quality data of young teachers in universities are collected and classified according to the influencing factors to obtain learning samples. The network prediction model is established,the sub network model is constructed between the primary and secondary indicators of teaching quality evaluation,and the deep neural network is constructed between the sub network output and the teaching quality prediction output. The hyper parameters of the network are adjusted by means of the genetic optimization algorithm,so that the network prediction performance is improved significantly. The simulation results show that,in comparison with the existing network model,the proposed deep neural network evaluation method is more accurate in predicting the teaching quality of young teachers in universities,and has a wide range of teaching management application value.