New Development Of K-means Algorithm

This paper presents a framework that combines the advantages of convolutional neural networks,unsupervised learning,and K-means clustering algorithms.When there is less available label data,the K-means algorithm and the convolutional neural network are combined to form a convolutional K-means algorithm.It can test and classify data more efficiently,and improve the accuracy of classification while ensuring the convergence speed.Through the learning of convolutional neural networks and cluster analysis,as well as the continuous optimization and adjustment of the network.The author uses the convolutional K-means algorithm to train deep convolutional networks.Because the task of training deep neural networks with labeled data is arduous and cumbersome,a large amount of labeled data is required to obtain the latest results.The layered function of unsupervised learning technology can be used to reduce the dependence on large amounts of labeled data,which proves that learning deep neural networks,The connections between the various layers of the network play an important role in improving the performance of unsupervised technology.Through experimental analysis,it is found that the supervised algorithm can perform better data clustering with fewer filters.As the number of filters increases,the supervised algorithm loses its accuracy and does not match the training set.The unsupervised learning algorithm convolution K-means algorithm does not work well when there are few filters.As the number of filters increases,the unsupervised learning algorithm can better represent data clustering and can improve the accuracy of test classification.Experimental results show that the convolutional K-means algorithm has a classification accuracy of 74.1% in the STL-10 dataset.This result shows that the convolutional K-means algorithm is superior to other unsupervised learning filter methods in the STL-10 dataset.