Open shlid007 opened 1 year ago
I have a similar error this is error sentence.
Warning messages: 1: In assign_score(validation_set_2, score_table) : NAs introduced by coercion 2: In print_roc_performance(y_validation, validation_set_3$total_score, : NA in the score: 1148
May I check what the sample size of your data is? Does it contain missing values? Is it an error or a warning? This is to say NA was introduced during dataset transformation. You can narrow it down to specific variables causing the problem. (e.g., delete one variable at a time)
Hi - sorry for late response; I am going to do more data preparation steps to see if that helps the error. I will let you know what I find! In meantime, I'm trying to understand the main advantages/difference with AutoScore compared to other interpretable models.
I realize parsimony is a defining attribute, but what about AutoScore's methodology makes it a parsimonious, optimal model? Is it how the random forest variable importance is baked in and variables iteratively selected based on AUC score?
Thanks for any feedback you can provide!
Sincerely, Sue
From: XIE FENG @.> Sent: Friday, June 23, 2023 9:26 PM To: nliulab/AutoScore @.> Cc: Sue Lhymn @.>; Author @.> Subject: Re: [nliulab/AutoScore] AUC step generates NA warning - significant? (Issue #7)
May I check what the sample size of your data is? Does it contain missing values? Is it an error or a warning? This is to say NA was introduced during dataset transformation. You can narrow it down to specific variables causing the problem. (e.g., delete one variable at a time)
— Reply to this email directly, view it on GitHubhttps://github.com/nliulab/AutoScore/issues/7#issuecomment-1605224186, or unsubscribehttps://github.com/notifications/unsubscribe-auth/A3MFT2NWXVUOOOXC5KWLLLTXMY63HANCNFSM6AAAAAAYLBQNPA. You are receiving this because you authored the thread.Message ID: @.***>
I am running the AUC step that calculates the parsimony plot. The log shows this warning and other similar ones - significant? I assume it is not a show-stopper since the number of NAs is relatively small compared to my dataset size.
Warning in compute_auc_val(train_set_1, validation_set_1, variable_list, : NA in the validation_set_2: 52