Correction method, device and terminal for searching text

Abstract

The embodiment of the invention discloses a correcting method, a device and a terminal for searching text. The method comprises the following steps: acquiring target searching text, and performing word segmentation processing on the target searching text to determine a searching term sequence corresponding to the target searching text; each candidate accurate word sequence corresponding to the search term sequence being determined according to the search corpus, wherein, the search terms in the search term sequence correspond to the candidate accurate words in the candidate accurate word sequence one by one; according to the searching corpus, searching word sequence, each candidate accurate word sequence, Wilson confidence interval formula and the total number of searching words, the candidate correction probability corresponding to each candidate accurate word sequence being determined; the target accurate word sequence corresponding to the target search text being determined according to the candidate correction probability, and the target accurate text corresponding to the target search text being determined according to the target accurate word sequence. Through the technical proposal of the embodiment of the invention, the situation that the calculated correction probability is zero can be effectively avoided, thereby improving the accuracy of correction.

Description

technical field [0001] Embodiments of the present invention relate to information processing technologies, and in particular to a method, device and terminal for correcting searched text. Background technique [0002] With the rapid development of information technology, users can search for information to obtain the required information. Usually, users can input search text according to their own needs, and find out the search results corresponding to the search text from the information collection by means of a search tool. For example, on a live webcast platform, users can enter the name of the host in the search entry, so that they can quickly find the live video they want to watch. [0003] Usually, users often make mistakes when inputting search text, such as spelling mistakes, reversed words, etc., so that users cannot find the desired search results, so the search text input by users needs to be corrected. [0004] In the prior art, the hidden Markov model is used ...

AI Technical Summary

Problems solved by technology

However, when the search text is misspelled, it is very easy to cause the state transition probability in the hidden Markov model to be calculated as zero, which in turn leads to the calculated correction probability as zero, but the actual correction probability is not zero

It can be seen that if the correction probabilities corresponding to multiple candidate accurate texts are all zero probability, the accuracy of the optimal accurate text corresponding to the search text determined by comparing the correction probabilities is low, thereby reducing the accuracy of the search results and affecting user search experience

Images