search

yandex-research / rtdl-num-embeddings

(NeurIPS 2022) On Embeddings for Numerical Features in Tabular Deep Learning
https://arxiv.org/abs/2203.05556
MIT License
312 stars 34 forks source link

How to inverse normalize the predictions ? #8

Closed manmeet3591 closed 1 year ago

manmeet3591 commented 1 year ago

I have run the code and have generated predictions.npz, the normalization I used was quantile with the toml file similar to california data performing regression. Can you please help me in inverse normalizing the predictions

Yura52 commented 1 year ago

If I understand your question correctly, you need to load y_info.pickle (it is saved in the same output directory) and:

  1. multiply the predictions by y_info['std']
  2. add y_info['mean'] to the predictions

Does this help?

manmeet3591 commented 1 year ago

Thank you so much. It solves my problem.