Non-semantic and non-phrase switching methods in song search

Abstract

The invention discloses a non-semantic non-word-group switching method in song search and provides an information processing method and device. The information processing method and device can be used for conducting word segmentation on information. The information processing method is a word segmentation method based on unitary word segmentation, binary word segmentation and dictionary word segmentation, wherein the unitary word segmentation, the binary word segmentation and the dictionary word segmentation are fused rather than mechanically superposed so as to form a new compound word segmentation method. Particularly in the song information search field, in the executing process of the method, corresponding judgments are made according to specific conditions, the characteristics of the song information search field and the characteristics of search strings of users, different word segmentation and search schemes are used according to different conditions, adoption of a complex semantic word segmentation scheme and other word segmentation schemes is avoided, and therefore search efficiency and accuracy are greatly improved.

Description

technical field [0001] The present application relates to an information processing method and an information processing device, in particular to a word segmentation processing method and device for performing double word segmentation on information in the field of information search, especially to non-semantic and non-phrase word segmentation in song search method and apparatus. Background technique [0002] In the field of information search, a very important factor affecting the quality of search results is word segmentation technology. That is, when processing the search material and the user's query statement, how to segment the text into multiple independent words, search for each word and then merge the results, which is related to the efficiency and accuracy of the retrieval. [0003] When Chinese word segmentation is carried out, because Chinese words are composed of single characters, and the same characters have different combined meanings in different contexts, ...

AI Technical Summary

Problems solved by technology

However, unary word segmentation will cause the document list of each word to be too long, and the overhead is very high when merging data

[0007] For binary word segmentation, binary word segmentation is a scheme proposed on the basis of unary word segmentation, which can effectively reduce the length of the document list of words and improve the efficiency of search, but it is poorer in fault tolerance processing and has a lower recall rate.

Images