Open AccessResearch and implementation of convolution k-means algorithm based on unsupervised learning
Author(s) -
Suyu Huang
Publication year - 2021
Publication title -
journal of physics. conference series
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.21
H-Index - 85
eISSN - 1742-6596
pISSN - 1742-6588
DOI - 10.1088/1742-6596/1883/1/012049
Subject(s) - unsupervised learning , computer science , cluster analysis , artificial intelligence , convolutional neural network , convolution (computer science) , pattern recognition (psychology) , wake sleep algorithm , deep learning , semi supervised learning , algorithm , artificial neural network , machine learning , generalization error
This paper proposes a framework combining the advantages of unsupervised learning, convolutional neural network and K-means clustering algorithm. When there are few label data available, convolution k-means algorithm is formed by combining k-means algorithm with convolution neural network. Through the learning of convolutional neural network and clustering analysis, as well as the continuous optimization and adjustment of the network, the convolutional k-means algorithm is used to train the deep convolutional network, and the hierarchical function of unsupervised learning technology is used to reduce the dependence on a large number of labeled data. Through the analysis of the experimental results, compared with the supervised learning convolution k-means algorithm, the unsupervised learning convolution k-means algorithm can better represent the data clustering with the increase of the number of filters, and can improve the accuracy of test classification. The unsupervised learning convolution k-means algorithm is better than other unsupervised learning filter methods.