Bootstrapping and Benchmarking Hyperparameters (5.5)

Bootstrapping is a method to average the scores from several runs of a keras neural network to average out some of the effects of the random nature of neural network weight initialization and training.  This video shows several techniques for benchmarking keras neural networks with an eye towards hyper-parameter optimization.

Code for This Video:
https://github.com/jeffheaton/t81_558...
Course Homepage: https://sites.wustl.edu/jeffheaton/t8...

Follow Me/Subscribe:
heatonresearch
https://github.com/jeffheaton
Twitter: jeffheaton

Support Me on Patreon: Patreon: jeffheaton

• #science_technology
• #python
• #python_introduction
• #python_introduction_deep_learning
• #deep_learning
• #neural_network
• #deep_neural_network
• #jeff_heaton
• #keras
• #tensorflow
• #washington_university
• #wustl
• #rmse
• #log_loss
• #ridge_regression
• #lasso_regression
• #regularization
• #bootstrapping_keras
• #bootstrapping