Multi-source data fusion method and system based on word vector matrix decomposition technology

Abstract

The invention provides a multi-source data fusion method and system based on a word vector matrix factorization technology, and the method comprises the steps: obtaining multi-source data samples, each multi-source data sample comprises multi-modal data of texts, voices, images and videos, and obtaining a corresponding implicit semantic knowledge base; jointly projecting the multi-source data samples and corresponding semantic information extracted from the implicit semantic knowledge base into a shared semantic subspace to generate a word vector matrix; decomposing the word vector matrix to obtain low-dimensional features of the multi-source data sample; training a classifier by taking the low-dimensional features of the multi-source data samples as input and taking corresponding semantic information as labels; multi-modal data of a to-be-mined target task is subjected to same word vector matrix generation and decomposition matrix processing and then input to the trained classifier, semantic information of the target task is obtained, and hidden semantic mining of the target task is completed. According to the method, the implicit meaning of the multi-source data and the relationship between the data are effectively extracted.

Description

technical field [0001] The present invention relates to the technical field of multi-source data fusion, in particular to a multi-source data fusion method and system based on word vector matrix decomposition technology. Background technique [0002] New tools for data processing free data scientists from the tedious work of data preparation, but how to customize and integrate multi-source data to form an effective analysis data set according to each data analysis project is still a more difficult problem that data scientists must face. Challenging bottleneck. [0003] With the development of computer technology and communication technology, and the continuous emergence of new theories and methods, multi-source data fusion technology will become more mature, from theoretical research to practical and wider application, and eventually to the direction of intelligence and real-time develop. Traditional word vector learning methods often rely on a large number of unlabeled te...

AI Technical Summary

Problems solved by technology

[0005] In view of this, the present invention proposes a multi-source data fusion method based on word vector matrix decomposition technology, which is used to solve the problem that the implicit meaning of multi-source data and the relationship between data cannot be effectively extracted

Images