Model evaluation IQRlogisticRegression projects

Two methods of model evaluation are applied to an IQRlogisticRegression object:

  1. Hosmer-Lemeshow goodness of fit test.

  2. Area under the receiver operating characteristic (ROC) curve. In order to consider that the logistic regression model developed has sufficient predictive performance, one could usually accept models with an AUC ROC>0.75 and a p value>0.05 for the Hosmer-Lemeshow test. Of course, personally preferred different values would be possible to define.

modelEvaluation_IQRlogisticRegression(
  project,
  filename = NULL,
  SIGNIF = 3,
  AUCthreshold = 0.75,
  HLpValueThreshold = 0.05
)

Arguments

project

Path to the IQRlogisticRegression to conduct the model evaluation for.

filename

File name (with path) to store the resulting figure.

SIGNIF

Number of significant digits used for the displayed p-value.

AUCthreshold

Threshold for acceptance of model (calculated value should lie >=)

HLpValueThreshold

Threshold for acceptance of model (calculated value should lie >=)

Value

IQRoutputFigure object with all results.

Details

Some references:

  • Hosmer-Lemeshow goodness of fit test

    • Hosmer 2000: Hosmer Jr. D, Lemeshow S. Applied Logistic Regression. 2nd ed. New York, NY: John Wiley & Sons; 2000.

    • Rana 2010: Rana S, Midi H, Sarkar S. Validation and performance analysis of binary logistic regression model. Paper presented at: The WSEAS International Conference on Environment, Medicine, and Health Sciences. 2010.

  • ROC

    • Hanley 1982: Hanley JA, McNeil BJ. The meaning and use of the area under a receiver operating characteristic (ROC) curve. Radiology. 1982;143(1):29-36.

    • Eberhart 2000: Eberhart LH, Hogel J, Seeling W, Staack AM, Geldner G, Georgieff M. Evaluation of three risk scores to predict postoperative nausea and vomiting. Acta Anaesthesiol Scand. 2000;44(4):480-488