This strategy consists of testing a predefined number of randomly sampled hyperparameter tuples. Sampling is done uniform at random within specified box constraints.
This solver implements the search strategy described in [RAND].
|[RAND]||Bergstra, James, and Yoshua Bengio. Random search for hyper-parameter optimization. The Journal of Machine Learning Research 13.1 (2012): 281-305|