Multi-label classification method for images based on gravity model

Abstract

The present invention relates to the field of machine learning, in particular to a multi-label classification method based on a gravitational model, comprising: obtaining a labeled sample set as a training sample set; calculating and sorting the distance between the training sample and other training samples to obtain the training sample Neighbor set; in the neighbor set, the positive correlation matrix is ​​constructed with the positive correlation between the labels, and the negative correlation matrix is ​​constructed with the negative correlation between the labels; the neighbor set of the sample to be tested is calculated, and the positive correlation matrix to be tested is constructed according to the neighbor set. The correlation matrix and the negative correlation matrix to be tested; the positive correlation matrix and the negative correlation matrix to be tested are used to obtain positively correlated data particles and negatively correlated data particles; a gravity model is constructed, and the positively correlated data particles and negatively correlated data are obtained through the sample to be tested The gravitational relationship between particles is classified; the invention introduces the consideration of the negative correlation between labels, fully utilizes the correlation between labels, and discovers the correlation in the neighbor set, avoiding global calculation and reducing the complexity.

Description

technical field [0001] The invention relates to the field of machine learning, in particular to a method for multi-label classification of images based on a gravity model. Background technique [0002] In the field of machine learning, classification problems occupy a large proportion. Traditional machine learning is mainly based on binary classification or multi-class classification, and its purpose is to accurately divide each data to be classified into a certain category. Such single-classification problems and multi-classification problems can be collectively referred to as single-label classification. In practical applications, most classification tasks need to face multi-label classification (multi-label classification) problems. For example, in a picture, the content of the picture may contain various elements, such as beaches, seas, tall buildings, people, and so on. Classifying such pictures is a multi-label classification task. [0003] The existing image multi...

AI Technical Summary

Problems solved by technology

However, the existing methods cannot make good use of label correlation or lead to a sharp increase in the complexity of classification methods in the exploration of label correlation.

Images