RandomProjection# class sdm.processing.numerical.RandomProjection(channels: int, *, init: Literal['normal'] = 'normal')# Bases: EnsembleProcessor Randomly project numerical columns. Parameters: channels (int) – Number of output numerical columns. init (Literal['normal']) – The projection weight initialization. "normal" samples Gaussian weights scaled by 1 / sqrt(channels). Capabilities# Handled Semantic Types numerical Requires Fitting ✅