A study on the assessment of learner ability based on the combination of auto-encoder and quadratic k-means clustering algorithm

At present, the widespread use of online education platforms has attracted the attention of more and more people. The application of AI technology in online education platform makes multidimensional evaluation of students’ ability become the trend of intelligent education in the future. Currently, most existing studies are based on traditional statistical methods to rank and evaluate students’ achievement, but this will lead to problems of single type of data and the inability of intra-class evaluation. In order to solve above problems in traditional statistical methods, a multidimensional learning ability evaluation method is proposed in this paper, which is based on auto-encoder and quadratic K-means clustering algorithm. It will be applied to the domain of intelligence education to evaluate students’ multidimensional learning ability. First, this method uses auto-encoder (AE) to reconstruct the students’ learning behaviour features in order to improve the clustering effect, then performs k-means clustering twice on reconstruction data. By using clustering to address the issue that cannot be addressed within the category, it ranks and evaluates students. This research employs a real data set of a particular platform for comparative studies in order to assess the performance of this strategy on various data sets. The results of the experiments demonstrate that this method performs much better than both the conventional clustering algorithm and the PCA-based reconstruction clustering method.