Wrongly written character monitoring method and system

Abstract

The invention provides a wrongly written character monitoring method, which comprises the following steps that: constructing a wrongly written character lexicon; carrying out data collection on a target website to obtain website data; carrying out preprocessing, webpage analysis and denoising on the obtained website data to obtain text contents; carrying out word segmentation processing on the text contents to obtain independent words; on the basis of a wrongly written character character lexicon, constructing an AC (Aho-Corasick) automaton dictionary tree, and generating a cache; constructinga context analysis model; and according to an AC dictionary tree cache and the context analysis model, carrying out wrongly written character identification, and outputting a wrongly written character identification result. By use of the method, wrongly written character monitoring accuracy can be effectively improved.

Description

technical field [0001] The present invention relates to the technical field of typo identification, in particular to a typo monitoring method and system. Background technique [0002] At present, some government websites have several serious typos on the same page or multiple pages with serious typos, which has attracted the attention of the public media and seriously affected the image of the government. In response to this phenomenon, the first national government website census included serious typos in the inspection points of the "serious mistakes" indicator. [0003] The typo monitoring method is usually composed of a typo lexicon, a word segmentation technology, and a typo recognition model. [0004] Word segmentation technology is the premise and key to identify typos. There are many word segmentation methods in the prior art, among which the word segmentation method based on character string is more general because it is relatively simple. String-based word segme...

AI Technical Summary

Problems solved by technology

[0005] 1. The existing word segmentation system only uses one word segmentation method (forward maximum matching method or reverse maximum matching method) for word segmentation. The word segmentation process is relatively rough, resulting in inaccurate word segmentation results and reducing the accuracy of word segmentation;

[0006] 2. The existing word segmentation methods usually only involve word segmentation in a single field, and it is still impossible to effectively identify strings in multiple fields

Images