Fragmented Knowledge Clustering method based on SOM

Abstract

Fragmented learning is a way to acquire discrete knowledge in the state of discontinuous time and space. In the development of the Internet and related technologies, it has become the new normal of personalized learning. This article considers the comprehensive application of ontology theory and knowledge aggregation to the process of knowledge reconstruction, and at the same time incorporates information technologies such as machine learning and semantic analysis, to establish a new self-organizing neural network-based fragmented knowledge cluster aggregation method. Solve the problem that existing knowledge organization methods are not suitable for fragmented learning. First, establish the concept of knowledge meta-data and define meta-data. And build knowledge clusters based on metadata and strong associations. The aggregated knowledge clusters contain richer and more systematic information, have stronger non-deterministic information description capabilities, and have stronger scalability. Second, establish knowledge cluster associations. The rational connection of knowledge clusters is conducive to the fusion and evolution of information, and the joint action of multiple knowledge clusters is an effective way to evolve new knowledge. Third, the integration of the concept of knowledge ontology and the insufficiency of the interaction between the organic knowledge clusters make up for the inaccuracy and incompleteness of the fragmented knowledge itself. The simulation proves that this method provides a new theoretical support and reform ideas for fragmented learning methods, and has a good role in promoting education fairness, optimizing education methods, and disseminating education content.