Chinese text error correction method based on same or similar pinyin

Abstract

The invention provides a Chinese text error correction method based on the same or similar pinyin, and the method comprises the following steps: S1, carrying out the adjustment on the basis of a conventional ngram language model, and building a Chinese character structure language model with the granularity of a single Chinese character; S2, performing candidate processing on the statement to be corrected to generate a candidate sequence; s3, performing error detection on the candidate sequence based on the confusion set and an MAD algorithm to obtain a to-be-corrected statement candidate sequence; and S4, based on the maximum posterior probability of the Chinese character structure language model, decoding and outputting an error correction result by using a double-selection Viterbi algorithm. Compared with a traditional method, the word granularity accuracy is high, and the error correction speed is higher.

Description

technical field [0001] The invention relates to the technical field of text error correction, in particular to a Chinese text error correction method based on the same or similar pinyin. Background technique [0002] Text error correction is applicable to many fields, such as manual typing assistance: it can automatically check and prompt typos after the user enters. In this way, errors caused by negligence can be reduced, and the efficiency and quality of user input can be effectively improved. In the field of search error correction: For search interfaces such as e-commerce and search engines, users often input errors when searching. By analyzing the form and characteristics of search items, you can Automatically correct search items and prompt users, and then provide search results that better meet user needs, effectively shielding the impact of typos on users' real needs; in the field of speech recognition or robot dialogue: embed text error correction into the dialogue ...

AI Technical Summary

Problems solved by technology

[0003] Aiming at the problem that the establishment of the confusion set requires a lot of time and manual maintenance, high cost and inconvenient use, and the accuracy of the traditional method is low, the present invention proposes a Chinese text error correction method based on the same or similar pinyin, and establishes The Chinese character structure language model whose granularity is a single Chinese character uses the confusion set and MAD algorithm to detect errors in the candidate sequence, and uses the double-choice Viterbi algorithm to decode and output the error correction result

Images