A method, device and computer-readable storage medium for constructing a dictionary

Abstract

Embodiments of the present invention provide a method and device for constructing a dictionary. The method includes: acquiring a candidate word and the definition of the candidate word; selecting a feature word of the candidate word from the definition of the candidate word; The feature words of the word, through the preset classifier, get the initial judgment result of the candidate word; according to the feature words selected from the interpretation of each middle word in at least one middle word, through the classifier, get each middle word The judgment result of word, wherein, this at least one intermediate word comprises the N-level characteristic words of this candidate word; According to the initial judgment result of this candidate word and the judgment result of this at least one intermediate word, determine the final judgment result of this candidate word, the The final judgment result of the candidate word is used to indicate whether the candidate word can be added to the dictionary. Therefore, the accuracy of the dictionary can be improved.

Description

technical field [0001] Embodiments of the present invention relate to the field of natural language processing, and more specifically, relate to a method and device for constructing a dictionary. Background technique [0002] Dictionaries are key resources in the process of natural language processing. At present, most of the dictionaries are based on artificially constructed dictionaries, that is, dictionaries that are manually recognized in the corpus. However, the disadvantage of artificially constructed dictionaries is that the words in the dictionary are not perfect. , especially for the existing emerging new words on the Internet, the shortcomings of artificially constructed dictionaries are more obvious, and cannot well meet the practical application. [0003] In order to make the construction of the dictionary more perfect, the way of automatically constructing the dictionary is introduced. At present, a known method of constructing a dictionary is to search for the...

AI Technical Summary

Problems solved by technology

[0004] However, this method of constructing a dictionary based on BoW features only considers the frequency of occurrence of each feature word in the feature words corresponding to the word, and treats each feature word as an independent entity, which cannot well represent the Semantic information, for example, the type tendency of some words is related to its idiomatic usage, common collocation and other information, which cannot be reflected by the BoW features extracted in the interpretation. come more interference, reduce the accuracy of the dictionary

Images