In here, I used the MNIST data to demostrate tSNE and DBSCAN algorithms.

The main function imports the MNIST csv file and reduces the high dimensionality to 2-D. Then plots the t-SNE figure with labels and the t-SNE figure with DBSCAN clustering.

The file includes two functions, DR_Plot and DR_Plot_black.

DR_plot: Input sample labels, n x 2 embedded array, and the "filename" you want to save. The output would be the scatter figure with color label depending on samples.

DR_plot_black: Input sample n x 2 embedded array, and the "filename" you want to save. The output would be the scatter figure.

To import the MNIST csv file and output with the data matrix and sample labels.