Kaggler v0.9.4 is released with additional features for DAE as follows:
- In addition to the swap noise (
swap_prob), the Gaussian noise (
noise_std) and zero masking (
mask_prob) have been added to DAE to overcome overfitting.
- Stacked DAE is available through the
n_layerinput argument (see Figure 3. in Vincent et al. (2010), “Stacked Denoising Autoencoders” for reference).
For example, to build a stacking DAE with 3 pairs of encoder/decoder and all three types of noises, you can do:
from kaggler.preprocessing import DAE dae = DAE(cat_cols=cat_cols, num_cols=num_cols, n_layer=3, noise_std=.05, swap_prob=.2, masking_prob=.1) X = dae.fit_transform(pd.concat([trn, tst], axis=0))
If you’re using previous versions, please upgrade
pip install -U kaggler.
You can find Kaggle notebooks featured with
Kaggler’s DAE as follows:
- Kaggler DAE + AutoLGB Baseline: shows how to train a LightGBM model using
Kaggler’s DAE features and
AutoLGBmodel at the current TPS May competition
- DAE with 2 Lines of Code with Kaggler: shows how to use
Kaggler’s DAE with the previous TPS April competition.
Any feedbacks, suggestions, questions for the package are welcome.
Hope it helps! :)