RBF-NN based fusion model for e-learning achievement evaluation

The trend of using e-learning as a learning and teaching tool is widely adopted by numerous organizations. In order to enhance the e-learning efficiency, there are some advantages in e-learning system: (1) repeatable (learning), (2) timeless, (3) distanceless and (4) spaceless. Because ldquostudent-centeredrdquo instruction is likely to become the primary trend in education, the e-learning system should consider both of personalization and adaptability. By using the online examination, we can obtain the learning levels of students to adjust the learning schedule instantly for each one and build more adaptive e-learning system. But, the biases of assessments are assigned by teacher under un-controllable condition (i.e. tiredness, preference). To overcome the drawback, this paper proposes a fusion model to assign learning achievements based on RBF-NN (radial basis function-neural networks) for assisting teachers. Proposed model utilizes similarity threshold to remove inconsistent data and make our achievements evaluation more reliable. To verify our model, this paper collects e-learning online examination data to illustrate and compare the performance of proposed model with conventional RBF-NN model. The performance comparison results show that the proposed model outperforms the conventional RBF model.