Machine learning algorithms in the context of digital education

Abstract

In response to the problems such as complex feature dimensions, large errors in manual weighting, and ambiguous classification boundaries in the evaluation of digital literacy of college foreign language teachers, a digital literacy evaluation model based on machine learning algorithms is constructed. In this paper, the identification of teachers' digital literacy is described as a supervised classification task driven by multi-source heterogeneous data. A feature space is established around teaching platform behaviors, digital resource calls, online interaction records, and evaluation feedback information. After data cleaning, normalization, feature selection, and vectorization encoding, the model input is formed. The model layer introduces support vector machines, random forests, and deep neural networks for hierarchical identification, and combines the cross-entropy loss function and parameter optimization mechanism to complete training. Experimental results show that the model has good performance in terms of accuracy, recall rate, and F1 value, and can provide a computable technical path for intelligent evaluation of teachers' digital literacy.