Multi-clustering-based unsupervised feature selection method and system

Abstract

The invention relates to the technical field of artificial intelligence, in particular to a multi-clustering-based unsupervised feature selection method and system, and the method comprises the steps: obtaining a plurality of to-be-clustered data points in a high-dimensional space, and obtaining a feature matrix; embedding the feature matrix in the high-dimensional space into a low-dimensional space by using an equal metric mapping algorithm to obtain an embedded matrix in the low-dimensional space; normalizing the feature matrix to obtain a reference feature matrix; feature vectors are formed by features in each dimension in the embedded matrix, and a sparse coefficient vector of each feature vector in the embedded matrix is obtained through fitting of the reference feature matrix and the embedded matrix; and selecting the maximum value of the sparse coefficient of each feature in the sparse coefficient vector as a contribution value, selecting the features corresponding to a plurality of maximum contribution values according to the preset number of features required to be selected to form a new feature matrix, and clustering based on the new feature matrix, so that the method is low in calculation complexity and accurate in clustering.

Description

technical field [0001] The invention relates to the technical field of artificial intelligence, in particular to a method and system for unsupervised feature selection based on multi-clustering. Background technique [0002] Machine learning algorithms often become difficult to work with when faced with hundreds or thousands of high-dimensional data. The best solution to this problem is to collect more data or use feature engineering to process the data, but collecting data often costs a lot of manpower and material resources, and is often a costly option. The feature selection method in feature engineering is a very effective method for processing features. It selects the features that are most helpful to the task and deletes useless features, thereby reducing the dimension of the data to make the data easy to process and prevent the data from being overused. fit. In most feature selection tasks, labeled data is often difficult to obtain. This is due to the high cost of m...

AI Technical Summary

Problems solved by technology

like figure 1 As shown, the Euclidean distance is figure 1 The black straight line in (a), and the geodesic distance is the red curve in Figure 1(a). When dealing with hypersurfaces in high-dimensional space, if the hypersurface is close to a plane, its two-dimensional distance in low-dimensional space The distance between points is approximately the same as the Euclidean distance, and when the hypersurface of the high-dimensional space is curved, if the Euclidean distance is used to embed the data, a large error will occur

Images