setGpFitnessFunc

Note: This function is not supported by Community Edition. You can get a trial of Shark from DolphinDB official website.

Syntax

setGpFitnessFunc(engine, func)

Arguments

engine is the engine object returned by createGPLearnEngine.

func is the fitness function, which can be:

  • A FUNCTIONDEF scalar, specifying a built-in function supported by fitnessFunc of createGPLearnEngine. It must be 'mse', 'rmse', 'pearson', 'spearmanr', or 'mae'.

  • A user-defined function with a floating-point return value. It takes two input arguments, the first represents the factor calculation result, and the second represents the predicted value. Currently it does not support complex assignment, if or for statement. Only return statement can be used to return a combination of the supported fitness function and the training function. For example:
    def f(x, y){
      return mse(x+y,y)
    }

    A user-defined fitness function can be combined with a helper function (as shown below) to pre-process the data before calculation.

    Function Num of Args
    clip(X,Y,Z) 3
    zscore(x) 1
    mad(X, [useMedian=false]) 1
    med(x) 1
    mean(x) 1
    corr(X,Y) 2

Details

Set fitness function for the GPLearn engine.

Examples

Set fitness function for the GPLearn engine.

def f(x, y){
  return mse(x+y,y)
}
setGpFitnessFunc(engine,f)