Database Reference
In-Depth Information
Metrics for different tree depths
Maximum bins
Finally, we will perform our evaluation on the impact of setting the number of bins for the
decision tree. As with the tree depth, a larger number of bins should allow the model to
become more complex and might help performance with larger feature dimensions. After
a certain point, it is unlikely that it will help any more and might, in fact, hinder perform-
ance on the test set due to over-fitting:
params = [2, 4, 8, 16, 32, 64, 100]
metrics = [evaluate_dt(train_data_dt, test_data_dt, 5,
param) for param in params]
print params
print metrics
plot(params, metrics)
fig = matplotlib.pyplot.gcf()
