1
$\begingroup$

The data could be classified correctly by method like LogisticRegression, NearestNeigbors.

enter image description here

The dataset is quiet simple, dataset wdx

This is not the first time I encounted with this method which across many versions.

Tested on 13.3 on windows 10.

As you can see, ListPlot,SetAlphaChannel many interesting symbols appear in the error information. This error information seems appear in the final report stage, not training stage.

Some sliced dataset is OK to train without error.

enter image description here

Split problem dataset into 2 is OK. enter image description here

A random dataset is OK, so the problem is related with some certain data.

dataset=Thread[RandomReal[{-1,1},{2000,40}]->RandomInteger[40,2000]];
$\endgroup$
6
  • 4
    $\begingroup$ You shouldn't ask people to download and import some random mx file. It's impossible to tell if it has any malignant code inside of it, so it's a potential vector for malware. Please provide a simple reproducible example in your post. $\endgroup$ Commented Nov 21, 2023 at 15:41
  • $\begingroup$ @SjoerdSmit OK, Let me see, actually, I don't know a small dataset to reproduce this problem. Let me do more experiments. $\endgroup$ Commented Nov 21, 2023 at 15:42
  • 2
    $\begingroup$ In that case I assume there must be some corrupt examples in the dataset. You can try and find a minimal example using bisection (split the dataset in two recursively and see which half has the problem) or something like that. $\endgroup$ Commented Nov 21, 2023 at 15:48
  • 1
    $\begingroup$ @SjoerdSmit I update a dataset with .wdx format which might be more safe, and I found split the dataset into two dataset to train is OK... so might not the problem of corrupt examples. $\endgroup$ Commented Nov 21, 2023 at 16:02
  • 4
    $\begingroup$ I think it's an issue with the interactive progress reporting. Adding the option TrainingProgressReporting -> "ProgressIndicator" fixes things on my end. $\endgroup$
    – Greg Hurst
    Commented Nov 21, 2023 at 16:29

0

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Browse other questions tagged or ask your own question.