Nonnegative local coordinate factorization-based clustering method

Abstract

The invention discloses a nonnegative local coordinate factorization-based clustering method, which comprises the following steps that: (1) a sample characteristic matrix is built; (2) a low-dimensional sparse matrix is iteratively outputted; (3) and the low-dimensional sparse matrix is clustered and analyzed. A sparse coding concept is introduced into the nonnegative matrix factorization (NMF) process, nonnegative local coordinate factorization is undertaken on a high-dimensional sample characteristic matrix, so the factorized coefficient matrix is used as a low-dimensional expression of the high-dimensional sample characteristic matrix, and the low-dimensional matrix is clustered to analyze, so the clustering analysis is simple and valid; and at the same time, data after the dimensional reduction has good explanatory property. Compared with the dimensional reduction method of the prior art, the judgment capacity of the clustering analysis can be further improved.

Description

technical field [0001] The invention belongs to the technical field of data processing, and in particular relates to a clustering method based on non-negative local coordinate decomposition. Background technique [0002] Clustering is a common multivariate statistical analysis method in machine learning and data mining. It discusses a large number of samples and requires reasonable classification according to their respective characteristics. There is no model for reference or to follow, that is, in performed without prior knowledge. At present, as an effective means of data analysis, clustering methods are widely used in various fields: in business, cluster analysis is used to discover different customer groups, and characterize the characteristics of different customer groups through purchase patterns; In biology, cluster analysis is used to classify animals and plants and to classify genes to gain an understanding of the inherent structure of populations; in geography, c...

AI Technical Summary

Problems solved by technology

[0009] These traditional clustering methods have successfully solved the clustering problem of low-dimensional data, but due to the complexity of data in practical applications, they often fail when dealing with many high-dimensional data.

Because traditional clustering methods mainly encounter two problems when clustering high-dimensional data sets: (1) there are a large number of irrelevant attributes in high-dimensional data sets, making the possibility of clusters in all dimensions almost zero; (2) ) The curse of dimensionality brought by high dimensions makes the practicality of some clustering algorithms almost zero

[0014] PCA is a traditional and classic unsupervised dimensionality reduction method, which has been widely used in various applications. This method can effectively find out the main features of the data, but it cannot effectively extract the category features of the data; LDA is a supervised The dimensionality reduction method, although the effect is good, but this method requires a large amount of data containing label information as training data, so it is only suitable as a dimensionality reduction method for classification, not as a dimensionality reduction method for cluster analysis; NMF As a basic dimensionality reduction framework, the data obtained by dimensionality reduction has good interpretability and has become a hot spot at present, but the effect of cluster analysis after dimensionality reduction is not ideal, and the discriminative ability of cluster analysis is still low. There is room for improvement

Images