Grids

The usual way to find the best parameters is to create many different models with different combinations of parameters, and then examine the appropriate metrics. These experiments are referred to as Grid Experiments because it is driven by a grid of parameter combinations.

For example, if you want to try models with the Bins parameter having values of 10 and 20 and you want Top Columns to have values 5, 10, and 20 then you would create 6 model experiments to evaluate, one for each 2 X 3 combination:

When specifying these parameters, the values are separated by semicolons. For the above example, you would specify the Top Columns grid parameter as “5;10;20”. You can adjust these to fine tune your Grid. Be aware that the more values used the longer your Grid will take to run. A complex Grid may take several hours to complete since it is creating hundreds of Experiments.

ML Studio has two approaches to finding the best parameters based on Grid Experiments: Auto Tune and Exhaustive Grid Search.