Efficient and accurate noisy word processing method for text based on knowledge graph

Abstract

The invention discloses an efficient and accurate noisy word processing method for a text based on a knowledge graph, and the method comprises the following steps: S1, building a to-be-filtered word library, building a business knowledge graph, and adding various homophones of business related words; S2, adding a weight to each word; S3, performing word segmentation on the text through a word segmentation tool; S4, correcting the text homophones into service words through a service knowledge graph, and recording all the service words appearing in the text; S5, matching the corrected text witha filtering word, wherein the recorded service word is not influenced by filtering; S6, outputting the filtered text. According to the method, text data is filtered, the method is accurate and stable,flexible expansion and real-time modification can be achieved, interference of text polyphonic synonyms can be overcome, sentence meanings are well reserved while noisy words are filtered, and efficient and accurate noisy word filtering processing capacity is provided for a spoken language-to-text scene.

Description

technical field [0001] The invention relates to the technical field of knowledge graphs, in particular to an efficient and accurate noise word processing method based on knowledge graphs. Background technique [0002] Knowledge graph (Knowledge Graph), known as knowledge domain visualization or knowledge domain mapping map in the library and information industry, is a series of different graphics showing the knowledge development process and structural relationship, using visualization technology to describe knowledge resources and their carriers, mining , analysis, construction, drawing and display of knowledge and their interrelationships, knowledge map, is through the application of mathematics, graphics, information visualization technology, information science and other disciplines of theories and methods and metrology citation analysis, co-occurrence analysis and other methods, and use the visualized map to visually display the core structure, development history, fron...

AI Technical Summary

Problems solved by technology

[0003] However, the current efficient and accurate noise word processing method filters out sensitive words through a complete sensitive word lexicon and matches text words. If a sensitive word is found in the text, the word will be filtered out from the text and only focus on sensitive words. In the spoken environment, many Speech auxiliary words are noise words, and there is no noise word filtering ability for spoken language to text. Homophone synonyms are easy to be improperly handled and accidentally deleted or omitted during filtering, which will affect the meaning of the sentence.

Images