cal_AUC_ml_res - Githubissues

l-magnificence / Mime

Machine learning-based integration model with elegant performance

Other

72 stars 14 forks source link

cal_AUC_ml_res #22

Open weifuq opened 3 months ago

weifuq commented 3 months ago

all.auc.3y <- cal_AUC_ml_res(res.by.ML.Dev.Prog.Sig = res,train_data = mm[["Training_Dataset"]],

inputmatrix.list = mm,mode = 'all',AUC_time = 3,

auc_cal_method="KM") --- Data preprocessing --- [1] "Training_Dataset" [1] "Testing_Dataset1" [1] 2832.4 [1] "Testing_Dataset2" [1] 2022 [1] "Testing_Dataset3" [1] "RSF" Error in survdiff.fit(y, groups, strata.keep, rho) : There is only 1 group 是因为分不出来组吗？生存分析的时候可以出结果尝试了mode = 'double'也是一样的错误 all.auc.1y <- cal_AUC_ml_res(res.by.ML.Dev.Prog.Sig = res,train_data = mm[["Training_Dataset"]], inputmatrix.list = mm,mode = 'double',AUC_time = 1,double_ml1 = "StepCox[both]",double_ml2 ="SuperPC", auc_cal_method="KM")

l-magnificence commented 3 months ago

Can you show your riskscore based on RSF model?

weifuq commented 3 months ago

l-magnificence commented 3 months ago

Your OS.time is day unit?

weifuq commented 3 months ago

yes

weifuq commented 3 months ago

执行cal_AUC_ml_res函数的时候是不是要在list_train_vali_Data中加上分组信息？

l-magnificence commented 3 months ago

No need. No such problems have been encountered.

weifuq commented 3 months ago

all.auc.3y <- cal_AUC_ml_res(res.by.ML.Dev.Prog.Sig = res,train_data = mm[["Training_Dataset"]],

inputmatrix.list = mm,mode = 'all',AUC_time = 3,

auc_cal_method="KM") --- Data preprocessing --- [1] "Training_Dataset" [1] "Testing_Dataset1" [1] 2832.4 [1] "Testing_Dataset2" [1] 2022 [1] "Testing_Dataset3" [1] "RSF" Error in survdiff.fit(y, groups, strata.keep, rho) : There is only 1 group Error is why？

kongwh99 commented 2 months ago

I have encountered the same problem, can you solve this problem?

weifuq commented 2 months ago

No发自我的手机-------- 原始邮件 --------发件人： kongwh99 @.>日期： 2024年7月16日周二下午1:41收件人： l-magnificence/Mime @.>抄送： weifuq @.>, Author @.>主题： Re: [l-magnificence/Mime] cal_AUC_ml_res (Issue #22) image.png (view on web) I have encountered the same problem, can you solve this problem?

—Reply to this email directly, view it on GitHub, or unsubscribe.You are receiving this because you authored the thread.Message ID: @.***>

l-magnificence commented 2 months ago

Sorry, currently I don't know how to solve it due to no happening in my system. Anyone knows?

csu-lzj commented 2 months ago

I had the same problem,Can i calculate it like this? time_roc_train <- timeROC( T = train_risk$OS.time,
delta = train_risk$OS,
marker = train_risk$RS,
cause = 1,
weighting="marginal", times = c(365, 1095, 1825), ROC = TRUE, iid = TRUE
)

l-magnificence commented 2 months ago

Yes, you can calculate AUC yourself using the results output by ML.Dev.Prog.Sig. Like this:

 risk.survivalROC <- survivalROC(
                Stime = x$OS.time,
                status = x$OS,
                marker = x$RS,
                predict.time = 365 * 1, 
                method = "KM"
              )