Method and device for classifying named entities

Abstract

The invention provides a method and device for classifying named entities. The method comprises the following steps of: reading LOD (Levels of Detail) nodes corresponding to the named entities to be classified from a linking open data (LOD) data set; using the type attributes of the LOD nodes corresponding to the named entities to be classified as marking types of the named entities to be classified; reading candidate types; and calculating the possibility that the named entities to be classified belong to the candidate types according to the marking types. According to the technical scheme of the invention, the named entities can be helped to classify by using mass information existing on the network.

Description

technical field [0001] The present invention relates to the field of data processing, and more particularly, to a method and device for classifying named entities. Background technique [0002] Classification of named entities is a key step in many applications. A named entity is essentially a word with semantic meaning. For example, in an automated question answering system, it is necessary to determine whether the type of the candidate answer matches the type specified by the question. In the information extraction system, it is necessary to determine the type of named entity in order to prepare for the subsequent extraction process. [0003] Traditional automated classification systems are based on machine learning. Specifically, a series of named entities of known types are input to the automatic classification system, and each named entity corresponds to a feature vector. The automatic classification system obtains the corresponding relationship between feature vect...

AI Technical Summary

Problems solved by technology

[0005] The above methods need to manually determine the appropriate type for the named entity used for learning, which will bring a large workload

Images