The cell below plots the error function of the classification model given different values of the `radius` split parameter (as in the introductory notebook). Each time this cell is run it will select two different subsets of the data, and plot the accuracy function (number of correct predictions) for each set. Investigate how the plots vary when you change the value of the parameter `N` (which controls the size of the sampled data used to calculate the accuracy curves), and when you re-run the cell for the same value of `N`. What do you observe in general?
Questions to discuss:
* Do the two curves (in blue and orange) generally have the same or different peaks?
* Does the location (parameter value) of the peak vary at all with the sample size?
