Personalised learning resource online recommendation method based on multi-dimensional feature extraction

In order to optimise the effectiveness of resource recommendation and improve the coverage of personalised learning resource recommendation results, a personalised learning resource online recommendation method based on multidimensional feature extraction is proposed. Firstly, based on the feature expression and density parameters of user behaviour data, cluster the users. Secondly, extract users' time features, preference features, and learning resource features, and use feature matrices for efficient feature mining. Finally, the extracted personalised learning resource features are input into the self-organising maps (SOM) network, and through the resource scoring mechanism and similarity calculation process, recommendation prediction values are generated and sorted to form a personalised recommendation set. The experimental results show that this method can accurately provide resource solutions that meet user needs when the number of resources and users increase, and the recommendation coverage rate always remains above 90%.