Lucene-based wrongly written character query method

Abstract

The invention relates to a Lucene-based wrongly written character query method. A sentence of a queried text is subjected to word segmentation; a first word is selected; whether the first word is a single word or not is judged; if the first word is the single word, a similar voice table and a similar shape table are queried; a query result simset is returned according to the similar voice table and the similar shape table; the query result simset and a next word or a query result simset of the next word are subjected to Cartesian product to obtain a Cartesian product result; the result is used to be matched with all words in a dictionary; if the matching succeeds, an error correction result is returned and added to an error correction result set; if the error correction result set is null, a null value is returned and the matching exits; if the error correction result set is not null, all error correction results are returned and used for performing a query; and if the first word in the sentence of the text is not the single word through the query or the matching between the result and all the words in the dictionary fails, characters are read backwards, and the previous steps are repeated. The method has the advantages that Lucene retrieval can be more accurate and humanized; and the retrieval accuracy is improved.

Description

Technical field [0001] The invention belongs to natural language processing in the field of artificial intelligence computers, and particularly relates to a query method based on Lucene typos. Background technique [0002] With the rapid development of information processing technology and the Internet, traditional text work has been almost completely replaced by computers. Text electronic publications such as e-books, e-newspapers, e-mails, and office documents continue to emerge, and there are more and more errors in the text. . [0003] At present, most of the manual proofreading methods are adopted. The proofreading work is monotonous, labor-intensive, and low in efficiency. The manual proofreading method can no longer meet the needs of text proofreading. Therefore, the study of automatic text proofreading has far-reaching significance for theory and application. Automatic text proofreading is one of the main applications of natural language processing, and it is also a diffic...

AI Technical Summary

Problems solved by technology

[0006] 3) True word errors will interfere with the grammar and semantics of the entire sentence, so finding true word errors requires a lot of knowledge and resources;

[0007] 4) Data sparseness is a major obstacle for automatic proofreading of true word errors

The automatic proofreading method for Chinese true word errors of the present invention solves the problems of data sparseness, misjudgment of correct words, and low proofreading efficiency in the prior art, and has high effectiveness and accuracy; but the inventive method still has certain defects: In practical applications, this method requires a large amount of corpus training, and the retrieval takes a lot of time, which is not very good for the actual user experience

Images