Laurae (version 0.0.0.9001)

LauraeML_utils.badlog: Laurae's Machine Learning Utility: bad input logger

Description

This function is a helper function to do the logging when a bad input is existing for LauraeML during a training iteration of the optimizer, specifically used when you have no features during training.

Usage

LauraeML_utils.badlog(mobile, logging, x, y, score = NA)

Arguments

mobile

Type: environment. The environment passed from LauraeML.

logging

Type: character. The logging parameter passed from LauraeML (where to store log file).

x

Type: vector (numeric). The hyperparameters to use passed from the trainer.

y

Type: vector (numeric). The features to use, as binary format (0 for not using, 1 for using) passed from the trainer.

score

Type: numeric. The score to optimize.

Examples

## Not run: ------------------------------------
# # What if we have no feature selected?
# if (sum(y) == 0) {
# 
#   # Logging specific
#   LauraeML_utils.badlog(logging, x, y, score = NA)
#   
#   # Last, we return an absurd score which is so high
#   # you would rather have a random model than this 0-feature model
#   # 
#   # This iteration will be ignored by the optimizer if
#   # it does not belong to the elite proportion of the optimization iteration
#   # which should be obviously true
#   return(LauraeML_utils.badinput(maximize,
#                                  score = 9999999))
# }
## ---------------------------------------------