Differences

This shows you the differences between two versions of the page.

--- projects:workgroups:patient-level_prediction:best-practice [2016/05/03 19:01] – [Best practices] prijnbeek
+++ projects:workgroups:patient-level_prediction:best-practice [2016/05/04 15:43] (current) – [Best practices] prijnbeek
@@ Line 11: / Line 11: @@
 ===== Best practices =====
-**Data characterisation and cleaning**: Before modelling it is important to characterize the cohorts, for example by looking at the prevalence of certain covariates. Tools are being developed in the community to facilitate this.
+**Data characterisation and cleaning**: Before modelling it is important to characterize the cohorts, for example by looking at the prevalence of certain covariates. Tools are being developed in the community to facilitate this. A data cleaning step is recommended, e.g. remove outliers in lab values.
 **Dealing with missing values **: A best practice still needs to established.
@@ Line 17: / Line 17: @@
 **Feature construction and selection**: Both feature construction and selection should be completely transparent using a standardised approach to be able repeat the modelling but also to enable application of the model on unseen data.
-**Inclusion and exclusion criteria** should be made explicit. It is recommended to do sensitivity analyses not he choices made. Visualisation tools could help and this will be further explored in the WG.
+**Inclusion and exclusion criteria** should be made explicit. It is recommended to do sensitivity analyses on the choices made. Visualisation tools could help and this will be further explored in the WG.
 **Model development** is done using a split-sample approach. The percentage used for training could depend on the number of cases, but as a rule of thumb 80/20 split is recommended. Hyper-parameter training should only be done on the training set.
-**Model validation** is done only once on the holdout set. The following performance measures should be added: To Do!
+**Internal validation** is done only once on the holdout set. The following performance measures should be calculated:
+  . Overall performance: Brier score (unscaled/scaled)
+  . Discrimination: Area under the ROC curve (AUC)
+  . Calibration: Intercept + Gradient of the line fit on the observed vs predicted probabilities
+We recommend box plots of the predicted probabilities for the outcome vs non-outcome people, the ROC plot and a scatter plot of the observed vs predicted probabilities with the line fit to that data and the line x=y added.