In my last post I was able to successfully re-orient a set of pixel data to reconstruct images of handwritten digits. SInce version 12 of JMP we have been able to create expression table columns that can contain images. That’s a logical location to store my newly revealed images:
So I have an image, and the value of the digit that it represents. What can I do with this data? It is a popular dataset used for testing machine-learning algorithms; so I thought it would be interesting to use the data to investigate some approaches to building predictive models.
Fundamentally the data is presenting me with a classification challenge: one that I want to simplify. So I’m going to take a subset of the data that represents only the digits 5 and 6:
Now the challenge is to build a model that differentiates between fives and sixes. I should be able to apply a number of different techniques to the data: logistic regression, neural networks, recursive partitioning, cluster analysis. Should be fun!