COBOL is still not dead.
Prediction with Decision Trees and Student Performance Data
Since the code is not available as part of the humble bundle download I decided to code along with the video. I found that I needed to install additional packages including pandas,pydot,graphviz and matplotlib. The graphviz needed to be installed under ubuntu and then use pip to install for python. I also had to figure out how to get the output to show when not using jupyter (http://jupyter.org/index.html)
It was a 6 and a half minute video that took me a few hours to go through with the tracking down and adding the packages. I don’t know how this was aimed at beginners.
Here’s a screen print of the final program. It used the pydot and graphviz to generate the decision tree graphic and then matplotlib to generate the graph of decision levels. The graph is from what I understand showing that the three points of data yields the best prediction.
I watched another Great Courses lecture on Writing Fiction.