Where communities thrive

  • Join over 1.5M+ people
  • Join over 100K+ communities
  • Free without limits
  • Create your own community
Repo info
  • Oct 28 03:19
    mjijas commented #248
  • Oct 28 01:42
    rasbt commented #248
  • Oct 28 01:41
    rasbt commented #248
  • Oct 28 01:35
    mjijas commented #248
  • Oct 17 15:02
    rasbt commented #742
  • Oct 17 15:02
    rasbt commented #742
  • Oct 13 16:53
    rasbt closed #741
  • Oct 13 16:31
    rasbt opened #741
  • Oct 13 04:31
    rasbt closed #740
  • Oct 13 04:29
    coveralls commented #740
  • Oct 13 04:14
    rasbt synchronize #740
  • Oct 13 04:03
    review-notebook-app[bot] commented #740
  • Oct 13 04:03
    rasbt opened #740
  • Oct 12 20:12
    k3ybladewielder commented #496
  • Oct 12 19:35
    k3ybladewielder commented #496
  • Oct 12 17:43
    rasbt closed #739
  • Oct 12 17:43
    coveralls commented #739
  • Oct 12 17:27
    review-notebook-app[bot] commented #739
  • Oct 12 17:27
    rasbt opened #739
  • Oct 12 17:22
    rasbt commented #738
Ali Ahmad Faruqi
Vijay Krishnavanshi
Ali Ahmad Faruqi
New to Gitter, so just testing it out! :)
Anyone got any experience on multiprocessing with SFS and gridsearch on an ESXI host. I am running into the wierdest issues
Sebastian Raschka
Sorry, I have no experience with this particular setup, but maybe it could be related to multiprocessing. The SFS does multiprocessing for CV. Have you tried setting n_jobs=1 for debugging the issue?
John C Vallier
I need help with the plot_decision_regions function, as I have a random forest classifier with 6 features. I am asked to provide filler values or sometimes i get X must be an integer found int64. How do I handle more than 2 features for this function? I looked over the documentation but still confused. By the way just want to thank you. Your books and mlxtend is simply the bomb.
Bhupendra singh
I'm using fpgrowth for finding frequent itemset, is there any way to make it faster using multiprocessing.
i'm using support value 0.005 and data can have a high number of transactions
Hello! New here, just discovered mlxtend while I am doing research for my Msc thesis. I was wondering, can Mlxtend be used for unsupervised feature selection? Cheers
Kev. Noel
@rasbt :
Have started to do a package metric, ONLY focus on metric evaluation cross platform, cross models.
Interface is functional (ie numpy style).
Reason is we spend time to re-implement the metrics many times....
What do you think ?
Archie Norman
Hi - I was wondering if anyone had any suggestions when using the apriori algorithm with collaborative filtering. Let’s say I have a user set with a list of tags that the user is in interested in. I’d like to generate recommendations based on the association rule if the product matches a category/tag
Sanket Shahane
Hi folks, I have a question about scalability tests of the association mining algorithms. Do you have any test reports I can go through? My data is going to be 100's of GB and roughly 100 million transaction
@rasbt ^
Sebastian Khan

Hi, I am interested in using the "Sequential Feature Selector" (SFS) to fit some N-dimensional data with one output variable where N could be up to 7.

Is it correct that the I can use the mlxtend's SFS algorithms to find the "best" set of basis functions for each dimension (found by fitting against some training data and then compared against a validation set to prevent overfitting?)

Also maybe it's not appropriate to post here but rather on the google group, just let me know and I can move over to there.

Thanks in advance!

Kev. Noel
State of Art model Zoo, cross-platform tensorflow, pytorch, Gluon, .... :
KV Subbaiah Setty
hello, this is KVS Setty,
KV Subbaiah Setty
hello, I am a trainer, teacher and a writer on Machine Learning and Data Science. Recently came across this wonderful tool Mlxtend. Where can I find more step by step tutorials with explanations ,I mean more jupyter notebooks.