Text clustering method on basis of automatic threshold fish swarm algorithm

Abstract

The invention discloses a text clustering method on the basis of an automatic threshold fish swarm algorithm. The text clustering method includes computing a similarity matrix of feature vectors of texts, acquiring an initial equivalent partitioning threshold of each text by a corresponding row of elements of the similarity matrix, performing initial equivalent partitioning for the texts and determining an initial clustering number and an initial clustering center; and adopting the artificial fish swarm algorithm in a combination manner, updating the state of each artificial fish according to global optimal information and local optimal information, searching a global optimal clustering center and clustering initial clustering results again. The text clustering method has the advantages that the initial clustering number and the initial clustering center are acquired by a process for automatically acquiring the thresholds, the global optimal clustering center is searched by the aid of the artificial fish swarm algorithm, accordingly, shortcomings that the traditional clustering method is sensitive to initial values and only relies on local data characteristics and the like are overcome, and the text clustering accuracy and the text clustering intelligence can be improved.

Description

technical field [0001] The invention belongs to the technical field of text clustering, and more specifically relates to a text clustering method based on an automatic threshold fish swarm algorithm. Background technique [0002] The continuous growth of network information makes it more important to organize and manage massive text information and facilitate users to obtain useful information. Text information is mostly unstructured or semi-structured data, from which to discover potentially useful knowledge patterns, text clustering technology is a very important method. Because clustering does not require pre-category labels, text clustering has been widely studied and applied. Text clustering can be used as a preprocessing step in natural language processing applications such as multi-document automatic summarization. It can also mine the interest patterns of different users for information services such as information filtering and personalized recommendation. It can a...

AI Technical Summary

Problems solved by technology

The use of artificial fish swarm algorithm for clustering can overcome the drawbacks of traditional clustering algorithms that only rely on local data characteristics and are sensitive to initial values. with certain human factors

Images