Online learning platform based on AI adaptive learning and collaborative filtering algorithm

Abstract

With the development of information computing, online learning platforms have become increasingly popular, but there are still problems such as unsuitable learning resource recommendations. Therefore, this study proposes an online learning platform based on artificial intelligence adaptive learning and collaborative filtering algorithms, with innovation primarily reflected in the knowledge tracking model and multiple improvements to the collaborative filtering algorithm. The study also introduces knowledge acquisition/knowledge forgetting mechanisms and learning progress speed, significantly improving the accuracy of identifying learners’ knowledge state evolution. In the improvements to the collaborative filtering algorithm, the study first integrates user feature weights, then combines confidence factors with cosine similarity calculations, and finally introduces constraint factors to optimize neighborhood selection. The experiment shows that the maximum accuracy of knowledge state prediction of the improved model is 80.5%, which is 9.2 higher than the second best convolutional knowledge tracking model. The root mean square error and r² are −0.005 and 0.005 higher than the second best deep crossover network, respectively. When the constraint factor is 2, the improved collaborative filtering algorithm achieves the minimum mean absolute error of 0.784. The improved algorithm achieves the best performance when the amount of neighboring nodes is 30, with a recommendation accuracy of 80.6%. At the same time, the teaching effect of the lab class is significantly better than that of the control class. In difficult questions, the correct answer rate is 4.9% higher than that of the control class, and there is statistical significance between the samples (p<0.05). It can thus be concluded that researching methods to improve resource recommendations on existing online learning platforms can effectively enhance the suitability of resource recommendations and learning efficiency, promote educational equity and resource optimization, and reduce learning costs.