A state judgingmethod for applying short text classification to video

Abstract

The invention provides a state judging method for applying short text classification to video, which comprises the following steps: S1, preprocessing the original text in the video process, includingword segmentation and removing deactivation; S2, vectorizing the preprocessed text, and converting the short text in the original text into a mapping matrix; the mapping matrix and the kernel matrix with the weight value of 1 are convoluted to obtain the potential multi-scale candidate semantic units in the text; S3, find that nearest neighbor word representation of each candidate semantic unit, and calculating the Euclidean distance of all word representation vectors in the candidate semantic unit and the vector space. The invention has the beneficial effects that the short text classification based on the convolution neural network is applied to the judgment of the state in the video interview, and can be applied to the video interview, the intention of the candidate can be understood, the artificial intelligence interview can be realized, and the judgment of the state in the video in other fields can be expanded.

Description

technical field [0001] The invention relates to a state judging method, in particular to a state judging method applying short text classification to video. Background technique [0002] At present, employers and job seekers mostly conduct interviews through on-site or video conferences. In terms of the technology of the interview dialogue system, the text classification method has been applied in many scenarios after long-term research; short text analysis is a basic task in the field of natural language processing, which can effectively help users discover information from massive short text resources. Useful information; but the problem of short text classification started late, and there has been no general and effective method. The focus of short text classification is on feature processing and classification algorithms. [0003] The problems and defects of the existing interview dialogue system are as follows: [0004] Interview efficiency: On-site interviews require...

AI Technical Summary

Problems solved by technology

[0007] 1. Due to the fact that the short text provides few words, the effective information provided is limited, and keywords appear in different contexts to express different semantics, that is, there are data sparsity and semantic sensitivity problems in the semantic feature vector representation of short texts

[0008] 2. The short text information expansion technology based on the topic model can alleviate the impact of data sparsity to a certain extent, but the training of the topic model relies on a large-scale external auxiliary corpus, and the external auxiliary corpus requires the same short text data to be expanded. The collection maintains semantic consistency, making the collection process of the external auxiliary corpus time-consuming and labor-intensive

[0009] 3. The method of using search engines to expand short text content has a large time complexity and is difficult to adapt to the processing of massive data or online data

[0010] 4. The short text modeling method based on deep learning only uses limited context for semantic synthesis. Since the short text contains limited word co-occurrence information, it cannot effectively solve the problem of semantic sensitivity

Images