Applying t-SNE on MNIST dataset and visualizing the digit classes in 2 dimensions using matplotlib library
This project demonstrate, how the t-Distributed Stochastic Neighbor Embedding(t-SNE) algorithm converts the high dimensional data into 2 dimensional data and then data can be visualized in a 2 dimensional plan using matplotlib library.
In simple words, t-SNE is a non-linear dimensionality reduction technique in which algorithm identifies the clusters of similar data points with multiple features. More details on t-SNE can be found from the below link: http://www.jmlr.org/papers/volume9/vandermaaten08a/vandermaaten08a.pdf