Machine Learning with R Cookbook - Second Edition

How it works...

In this recipe, we demonstrate how different clustering methods work in regard to a handwriting dataset. The aim of the clustering is to separate 1 and 7 into different clusters. We perform different techniques to see how data is clustered in regard to the k-means and DBSCAN methods.

To generate the data, we use the Windows application paint.exe to create a PNG file with dimensions of 28 x 28 pixels. We then read the PNG data using the readPNG function and transform the read PNG data points into a scatter plot, which shows the handwriting digits in 17.

After the data is read, we perform clustering techniques on the handwriting digits. First, we perform k-means clustering, where k=2 on the dataset. Since k-means clustering ...