Retrieval Algorithm of Digital Information Resources for Legal Theory Teaching Based on Multi-scale Dense Network

To solve the problem that the existing retrieval algorithms have poor retrieval ability when retrieving digital information resources for legal theory teaching, the loss value in the iterative process of the retrieval algorithm is too large, which affects the retrieval performance and cannot meet the high-precision requirements of users for resource retrieval, the multi-scale dense network theory is introduced, Research on the design of digital information resources retrieval algorithm for legal theory teaching based on multi-scale dense network. Based on web crawler technology, obtain legal theory teaching document data sources, merge and transform different documents, introduce multi-scale dense networks, combine K-means algorithm and Canopy algorithm to cluster resources, allocate legal theory teaching digital information resources, and extract resource retrieval features based on word matching features, central sampling index library features, and topic related features, efficient retrieval of resources through information fusion. The experimental results show that the new retrieval algorithm has stronger retrieval ability in practical applications, and the loss value can rapidly decline in the algorithm iteration process, thus providing users with higher quality retrieval services.