Bias variance decomposition analysis

[vxbrandon]

• Fixes #122
• Compare the bias and variance of three models: Sklearn random forest, random forest with exact solver, and random forest with mab solver.
• Use the toy dataset where only the first three features of input data are relevant to the target data.
• Except setting num_trees=10 for sklearn and exact solver model and num_trees=20 for random forest with mab solver, equate other parameters of the models.

Result

• Run tests/bias_variance_analysis.py
• Then you will get the figure below.

[vxbrandon]

Hey @motiwari! Can you check if the new commit looks good? I made codes to write the numerical values in the plot!

[motiwari]

Thanks for the ping, yes looks good!