djstrouse / information-bottleneck Goto Github PK

When running the script "example_experiments" from a Jupyter notebook and then calling "test_IB_single()", the following error is raised in line 234 in IB.py:

ValueError: smoothing_type not yet set

Hello,

I read Tishby's article and roughly understand the algorithm and there was no mention of smoothing anywhere.
Could tell me what the smoothing parameter does ?
It might be usefull to include some infos about the dataset class in the readme.

Thanks !

Hello, I'd like to try your code on a dataset that has about 400 examples and 10 features per example.

I seem to have understood that your code assumes data with 2 features? Or can it work on data with N features?

Is there a way to use your implementation to treat 1D integer data ?
The way I understood it from the examples and the code, the input data needs to be 2D an continuous (I tried putting 2D integers as input, it didn't work).
Please correct me if I am wrong and redirect me to the corresponding function if I missed it.