Large Datasets
In practice, a large data set leads to a large K, and storing K may become a problem. One way to deal with this is to perform clustering on your large dataset, and populate the kernel with the means of those clusters. Since even this method may yield a relatively large K, it is common to compute only the top P eigenvalues and eigenvectors of K.
Read more about this topic: Kernel Principal Component Analysis
Famous quotes containing the word large:
“... when you make it a moral necessity for the young to dabble in all the subjects that the books on the top shelf are written about, you kill two very large birds with one stone: you satisfy precious curiosities, and you make them believe that they know as much about life as people who really know something. If college boys are solemnly advised to listen to lectures on prostitution, they will listen; and who is to blame if some time, in a less moral moment, they profit by their information?”
—Katharine Fullerton Gerould (1879–1944)