search

ohdsi-studies / PCE

Network study validating the Pooled Cohort Equation Model
3 stars 3 forks source link

Error in findrow #4

Open AEW0330 opened 3 years ago

AEW0330 commented 3 years ago

Here is what I see in the console:

Fetching cohorts from server Loading cohorts took 0.412 secs Fetching outcomes from server Loading outcomes took 0.236 secs Creating population Warning: no non-missing arguments to min; returning Inf Outcome is 0 or 1 Applying and evaluating model Starting Prediction 2020-10-27 07:01:39 for 2310 people Outcome count: 93 people Prediction completed at 2020-10-27 07:01:40 taking -0.094 secs Starting evaulation at 2020-10-27 07:01:40 AUC: 70.65 95% lower AUC: 65.08 95% upper AUC: 76.23 AUPRC: 9.42 Brier: 0.04 Calibration in large- Mean predicted risk 0.0361 : observed risk 0.0403 Weak calibration intercept: 0.1233 - gradient:0.9304 Hosmer-Lemeshow calibration gradient: 0.69 intercept: 0.02 Average Precision: 0.10 Evaluation completed at 2020-10-27 07:01:40 taking -0.513 secs Starting covariate summary at 2020-10-27 07:01:40 Covariate summary completed at 2020-10-27 07:01:40 taking -0.26 secs Error in findrow(fit, times, extend) : no points selected for one or more curves, consider using the extend argument

jreps commented 3 years ago

Seems to be an issue with the survival package https://rdrr.io/cran/survival/src/R/survfitms.R - I'll add a tryCatch around all the survival stuff so this doesn't stop the package - it will mean you might end up having some performance metric missing though.

Jake-Gillberg commented 3 years ago

BTW, these sorts of errors didn't appear to be fatal (there were tons of them in the run that just completed). The error that was causing the run to fail was insidiously, curiously, and probabalistically silent, but also appeared to be related to small sample sizes. (this worked: https://github.com/ohdsi-studies/PCE/pull/5/files#diff-03409357b835eb3982cfba82465d38dc11a2b3ccae1930db3c6a908bab9908acR232)

In a few of the failures I was blessed with an error that looked like this: ../nptl/pthread_mutex_lock.c:81: __pthread_mutex_lock: Assertion mutex->__data.__owner == 0' failed.