A neural network method for Cambodian entity recognition based on topic model word vector

Abstract

The invention relates to a neural network Cambodian entity recognition method based on a subject model word vector, belonging to the technical field of natural language processing. The invention firstly obtains the Cambodian text corpus and pretreats the corpus; and then constructs the theme model; the topic number of each word in the text is obtained by using the constructed topic model, and thetopic number is regarded as a pseudo word; putting the preprocessed text and the pseudo-words into the same corpus text, and use skip ram model to process word vector of each word and the subject vector corresponding to each word in the text are obtained by gram model processing. Cascading the word vector and the topic vector obtained in the step to obtain a subject term vector; finally, the obtained keyword vector is input as an input feature into the constructed depth learning model, and then the entity recognition of Cambodian is realized. The invention can better solve the problem of polysemy and homophone polysemy existing in the text, and has high recognition accuracy of Cambodian named entity.

Description

technical field [0001] The invention relates to a neural network Cambodian entity recognition method based on a topic model word vector, and belongs to the technical field of natural language processing. Background technique [0002] With the rapid development of the modern economy, exchanges and cooperation between my country and Southeast Asian countries have become more and more frequent, and exchanges and cooperation with Cambodia in economics, culture, education, etc. are also increasing year by year. In the context of the increasingly close development of China and Cambodia, it is particularly important to pay attention to and learn the cultural knowledge of Cambodia, but at the same time, the language barrier between the two countries has brought many difficulties to this task. Therefore, there is an increasing need to use natural language processing techniques to solve these difficulties. [0003] Cambodian language, also known as Khmer language, belongs to the Khme...

AI Technical Summary

Problems solved by technology

[0005] The present invention provides a neural network Khmer entity recognition method based on topic model word vectors, which is used to solve the problems of low recognition accuracy of Khmer named entities and polysemous and homonymous entities in Khmer entity recognition question

Images