importance – Datasets for Weighted Supervised Training

class pybrain.datasets.importance.ImportanceDataSet(indim, targetdim)

Bases: pybrain.datasets.sequential.SequentialDataSet

Allows setting an importance value for each of the targets of a sample.

addSample(inp, target, importance=None)

adds a new sample consisting of input, target and importance.

Parameters:

• inp – the input of the sample
• target – the target of the sample

Key importance:

the importance of the sample. If left None, the importance will be set to 1.0