Multi-label text classification algorithm based on hierarchical Transs-CNN

A text classification and multi-label technology, applied in text database clustering/classification, unstructured text data retrieval, semantic analysis, etc., can solve the lack of semantic information, the inability to fully capture text semantic information, and the inability to obtain sentences and sentences Questions like paragraphs vs. paragraph details

Pending Publication Date: 2021-10-29
YANSHAN UNIV
View PDF0 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0005] In view of the above-mentioned deficiencies, the purpose of the present invention is to solve the problems that the semantic information of the text cannot be fully captured, the detailed information between sentences and between paragraphs cannot be obtained, and the partial missing of semantic information is caused.

Method used

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View more

Image

Smart Image Click on the blue labels to locate them in the text.
Viewing Examples
Smart Image
  • Multi-label text classification algorithm based on hierarchical Transs-CNN
  • Multi-label text classification algorithm based on hierarchical Transs-CNN
  • Multi-label text classification algorithm based on hierarchical Transs-CNN

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0036] The technical solutions of the present invention will be clearly and completely described below through specific embodiments. Apparently, the described embodiments are only some of the embodiments of the present invention, but not all of them. Based on the embodiments of the present invention, all other embodiments obtained by persons of ordinary skill in the art without creative efforts fall within the protection scope of the present invention.

[0037] Such as figure 1 As shown, the multi-label text classification algorithm based on hierarchical Transformer-CNN in this embodiment includes the following steps:

[0038] S1, data preprocessing; use two very classic data sets for multi-label classification tasks, ReutersCorpus Volume I (RCV1) and AAPD, the former is the news field, which is Reuters' artificially labeled news data set, whose content comes from 1996 News from 1997 to 1997, the latter is mainly in the field of scientific papers, a large data set in the fie...

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More

PUM

No PUM Login to View More

Abstract

The invention relates to the technical field of natural language processing, in particular to a multi-label text classification algorithm based on a hierarchical Transs-CNN. The algorithm comprises the following steps: S1, preprocessing data ; S2, performing feature extraction on words; S3, performing feature extraction on sentences; S4, fusing the word features and the sentence features; Ss5, performing feature extraction on the fused features through a convolutional layer; and S6, classifying the texts through a full-connection network according to the obtained convolution features. The problems that text semantic information cannot be fully captured, detailed information between sentences and between paragraphs cannot be obtained, and semantic information is partially lost are solved.

Description

technical field [0001] The invention relates to the technical field of natural language processing, in particular to a multi-label text classification algorithm based on hierarchical Trans-CNN. Background technique [0002] With the rapid development of Internet information technology and the advent of the 5G era, massive data information is growing explosively, among which text data is the most abundant. Acquiring the core content of text information quickly and accurately in a limited time has become a practical challenge. Text classification is a cornerstone that can effectively solve the problem of information overload. By classifying text content with tags, it can effectively provide basic information for text retrieval and text recommendation, thereby greatly improving people's retrieval efficiency and reading experience. [0003] Text classification has always been the most basic but extremely important research field in natural language processing. Its goal is to au...

Claims

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More

Application Information

Patent Timeline
no application Login to View More
IPC IPC(8): G06F16/35G06F40/30G06N3/04
CPCG06F16/35G06F40/30G06N3/045
Inventor宫继兵王成龙房小涵
OwnerYANSHAN UNIV