AI again

 

 

Predicting Birds using Random Forests

I started the next video.  The dataset is from http://www.vision.caltech.edu/visipedia/CUB-200-2011.html.
The data files come in at around 1.3 GB. This has thousands of photos of birds. This video does not use the photos only the attributes. I spent a good bit of time trying to get the labels to show up properly on a confusion matrix but it was having none of it.

The left hand side is supposed to be the labels. They are all there just an overlapping mess.

You need to have a pretty beefy machine to run this stuff. I spent another good bit of time coding along with the video. Each time I add code it runs longer and much longer. The video was under ten minutes but the code has been running for over a half hour. Hopefully the last step that plots the output doesn’t have an error.

One thing I learned today that might come in handy is the pivot for dataframes in pandas.

Dataframe Pivot

There are times when you have a set of data by date that you would like to pivot and this looks like it would make that easy.