Regularized Fitness¶
Optimization ยท Fitness
A loss of a list of values against a target value: their mean absolute error (MAE) to the target, which measures accuracy, plus k times their sample standard deviation (STD), which measures consistency. The lower the score, the better the fitness.
Inputs¶
| Name | Type | Access | Default | Description |
|---|---|---|---|---|
values |
float | list | A list of numerical values to evaluate. | |
target |
float | item | The target value to compare against. | |
k |
float | item | 1.0 |
Weight of the STD in the loss. Optional. |
Outputs¶
| Name | Description |
|---|---|
fitness |
The fitness score, MAE + k STD; lower is better. 999 for no values. |