Performing GMRQ

[drdoppio]

At the stage of Gmrq.ipynb of your tutorial I have gotten the following errors:

Input In [1], in <cell line: 87>()
84 no_clusters=[700,800,900]
85 tica_lagtime=[2,4,6]
---> 87 GMRQ(no_of_features,no_of_components,no_clusters,tica_lagtime)

Input In [1], in GMRQ(no_of_features, no_of_components, no_clusters, tica_lagtime, parameter, gmrq_dir, clustering_dir)
56 else:
57 print('Running the {}th cycle'.format(i+1))
---> 58 cv = KFold(len(trajectories), n_folds=6, shuffle=True) #split the dataset into training set and test set
59 results = []
60 print_results = [] #train score, test score

TypeError: init() got an unexpected keyword argument 'n_folds'

Any suggestions?
thanks
Sandor

[drdoppio]

sklearn.cross_validation is not working
I have changed it to
sklearn.model_selection

[ykhdrew]

Hi Sandor

Were you able to solve the issue? I tried with a newly-created conda environment with my script i.e. conda env create -n msmbuilder -f environment.yml and GMRQ.ipynb works fine on my end.

This tutorial was using a very old scikit learn and things might be quite different in newer version.

[drdoppio]

I was not able to resolve the problem with changing the n_folders to n_splits I still have the error: Input In [1], in <cell line: 87>() 84 no_clusters=[700,800,900] 85 tica_lagtime=[2,4,6] ---> 87 GMRQ(no_of_features,no_of_components,no_clusters,tica_lagtime) Input In [1], in GMRQ(no_of_features, no_of_components, no_clusters, tica_lagtime, parameter, gmrq_dir, clustering_dir) 56 else: 57 print('Running the {}th cycle'.format(i+1)) ---> 58 cv = KFold(len(trajectories), n_splits=6, shuffle=True) #split the dataset into training set and test set 59 results = [] 60 print_results = [] #train score, test score TypeError: __init__() got multiple values for argument 'n_splits'

…

--

__________________________________________ Dr. Sandor Lovas, Professor Associate Editor, Frontiers in Chemical Biology Department of Biomedical Sciences Creighton University, School of Medicine 2500 California Plaza Omaha, NE 68178, USA Phone: 402-280-5753 Fax: 402-280-2690 On Mon, 2023-06-05 at 17:13 -0700, Andrew Yik wrote: Were you able to solve the issue? I tried with a newly-created conda environment with my script i.e. conda env create -n msmbuilder -f environment.yml and GMRQ.ipynb works fine on my end. This tutorial was using a very old scikit learn and things might be quite different in newer version. — Reply to this email directly, view it on GitHub<#1 (comment)>, or unsubscribe<https://github.com/notifications/unsubscribe-auth/A4CNJD5Z2C4YUN56722Q6QTXJZY4DANCNFSM6AAAAAAY3PR3GY>. You are receiving this because you authored the thread.Message ID: ***@***.***>

[msinclair-py]

KFold no longer takes data at the first position as in this tutorial. The KFold line should read as:
kf = KFold(n_splits=6, shuffle=True)

Then in the for loop:
for fold, (train_index, test_index) in enumerate( kf.split( list( range( len( trajectories ) ) ) ) ):