f1_scoreand the actual
average_depthof the model. We are interested in models that achieve higher than 0.8 of
average_depthlower than 10. It is also a good idea to store other metrics to inspect the models further. For example, we can keep track of each model's
inference_timeon the test set.
optimizestrategy with the
average_depth), all models failed to achieve low inference time. All-Constraint experiments recognize that other goals may exist, and they search for a diverse range of outcomes to service future demands. Specifically, note that:
inference_time. Notice that only models with low
num_boost_roundyields high F1 score -- this is not surprising, but our Multimetric experiment learns this and then spends its energy exploiting that information to make a better Pareto frontier.
num_boost_round. That is critical for producing models with good performance and faster inference time.
eta(learning rate) values between [-1.5, 0].
max_depththan Multimetric, especially between values 5 and 15.
min_child_weightvalues seems to produce acceptable results -- the metrics seem unaffected by this parameter alone. However, for satisfactory models, it looks like
min_child_weightare inversely correlated.