Showing Predictor "PBC_Classic_Visit1"

This predictor was successfully trained!

Meta

Notes:

Using 312-patient classic PBC survival dataset, which has follow-up data for patients (but for this experiment we're only using what is presumed to be their initial visit)

This data is being used by a group in Russ' PGM class, that is being coached by me

Publicly Accessible: yes

Viewable by Administrators: yes

Actions:Edit Delete

Hide

General Statistics
#Samples	312
#Censored	172 (55.1%) of 312
#Features	16 (16) = 16 Numerical + 0 Binary + 0 Nominal (0)

Dataset Show

Options
Regularization: L2	Tuning Score: avg log-likelihood loss
Time Points: 18	Tune Parameters: true
Feature Selection: false	Include Sibling Features: false
Standardize: true	Run Cross-Validation: true
Objective Function: log-likelihood	Marginal Loss Type: weighted
Use pre-defined Fold: No

Results
Start Time: 2018-11-30 16:35:38 -0700	Used Time: 1 minute
Label Name: time	C Parameter: 1.0
PSSP model: mtlr_model_file3127.mtlr	Version: 0.0.3

Predictor Settings / Retrain Show

Cross-validation results

5-Fold Cross-Validation Statistics*
Measure	PSSP Predictor (median)	PSSP Predictor (mean)	K-M Predictor
Concordance Index	0.83 ± 0.03	0.83 ± 0.03	0.5 ± 0.0
Hinged L1 Loss	589.0 ± 72.35		1008.3 ± 7.57
Uncensored L1 loss	1097.77 ± 120.14	1313.87 ± 106.88	1946.07 ± 42.2
Marginal L1 Loss	1051.33 ± 80.59	1129.94 ± 32.04	1815.32 ± 99.77
Marginal L2 Loss	1407.35 ± 142.8		2011.58 ± 94.32
Hinged L1 Log-Loss	0.34 ± 0.04	0.39 ± 0.02	0.54 ± 0.02
Uncensored L1 Log-Loss	0.69 ± 0.07	0.82 ± 0.05	1.13 ± 0.04
Marginal L1 Log-Loss	0.48 ± 0.04	0.54 ± 0.02	0.79 ± 0.04
Log-Likelihood Loss	2.13 ± 0.16
D-calibration χ² statistics	13.02 ± 3.68		0.73 ± 0.19
D-calibration χ² p-value	0.8 ± 0.2		1.0 ± 0.0

*Mean value ± standard deviation.

Download Predictions (CSV)

Individual Predictions

D-Calibration Histogram

Kaplan Meier Visualization

Show Feature Weights

Examine Classification Accuracy

PSSP

Showing Predictor "PBC_Classic_Visit1"