Research of Study Evaluation in E-learning System Based on UD and RBPNN
Jing Feng, Kang ShiYing · 2008
At present, the study evaluation in e-learning based on neural network is very little in China. The reason lies in the difficulty to find high quality training samples for self-learning and the training lacks strict scientific experimental design.In this paper, we have selected representative, uniformity and large-scale samples with uniform design (UD). And then use those samples to train the self-adaptive RBFNN which is applied to carry out the study evaluation in e-learning. The experiment shows that the generalization ability of self-adaptive RBFNN combined with UD has been greatly improved. The designed evaluation method realizes the self-adaptive, self-learning and non-linear approaching ability, meantime avoids the subjectivity and uncertainty of traditional evaluation.