Document clustering method based on distribution-convergence model

Abstract

The invention discloses a document clustering method based on a distribution-convergence model. The document clustering method comprises the steps of firstly, giving out a co-occurrence matrix establishing method based on the distribution-convergence model, counting co-occurrence frequency of knowledge property by utilizing the distribution-convergence model, and establishing a co-occurrence matrix by combining a hash map; secondly, establishing a nearness degree matrix by combining the co-occurrence matrix with a nearness degree coefficient; thirdly, standardizing the nearness degree matrix; finally, clustering the nearness degree matrix by using a sum of squares method, and realizing efficient fine-grit clustering on knowledge. The document clustering method disclosed by the invention can be applied to a network document knowledge base for document clustering, so that relatively ideal clustering accuracy and relatively ideal calculating efficiency can be achieved, fine-frit document clustering can be realized, and meanwhile, the time cost is reduced.

Description

technical field [0001] The invention relates to a knowledge clustering method, in particular to a document clustering method based on a distribution-convergence model, which belongs to the cross-technical application field of data mining and knowledge systems. Background technique [0002] Generally, the network document knowledge base system mainly classifies the stored documents according to the subject category, but does not carry out finer-grained classification according to the knowledge field under the subject. The coarse-grainedness of this classification makes learners prone to cognitive trek and knowledge overload in the process of document retrieval and reading. Integrating and counting knowledge objects according to their attributes through document clustering can not only classify documents in a more detailed manner, clarify the knowledge context for learners, improve the efficiency of document research, but also reveal the potential of knowledge development laws...

AI Technical Summary

Problems solved by technology

[0005] The technical problem to be solved by the present invention is: to provide a document clustering method based on the distribution-convergence model, and to use the distribution-convergence model to construct a co-occurrence matrix, which solves the problem that the limited memory of a single computing node is difficult to store and process a large matrix. Difficulties such as clustering or low clustering efficiency

Images