Method and device for text clustering and electronic device

Abstract

The invention relates to a method and device for text clustering and an electronic device. The method for text clustering comprises the steps of combining a plurality of original document sets of different themes into a document union set; arranging documents in the document union set in an ascending order, and obtaining an ascending order document union set; sequentially calculating similarity of the first document in the ascending order document union set and all document after the first document; if the similarity is larger than or equal to a first threshold, assigning the document and the first document to a class; if the similarity is smaller than the first threshold, marking the document as a non-classified document; sequentially executing similarity calculation and classification of the first document in the ascending order document union set and all non-classified documents after the first document. According to the method and device for text clustering and an electronic device, operations of word segmentation and feature extraction and the like are avoided, steps are simple, the accuracy rate is high, and the method and device for text clustering and the electronic device have language irrelevance and suitable for text clustering of various languages. Besides, the clustering speed and precision can be flexibly adjusted to meet different actual requirements.

Description

technical field [0001] The present invention relates to the technical field of text mining, in particular to a text clustering method, device and electronic equipment. Background technique [0002] As the name implies, clustering is the process of dividing the entire data set into several groups according to certain characteristics and rules. Elements within each group have high similarity in certain characteristics, while elements between groups have greater similarity in these characteristics. The resulting groups are a cluster, also often referred to as a "cluster". Currently, text clustering methods include partition clustering, hierarchical clustering, density-based clustering, semantic-based clustering, and clustering based on various model theories. [0003] Most of the above clustering methods need word segmentation or feature item support, so feature selection or dimensionality reduction is an important research content. For languages ​​such as Chinese and Uyghur,...

AI Technical Summary

Problems solved by technology

For languages ​​such as Chinese and Uyghur, the support of word segmentation is often inseparable, and the corresponding accuracy and speed of word segmentation will also follow, which will eventually affect the speed, accuracy and recall of clustering

Images