Page tree


: Model Validation = Regression Validation

Validation of the Model is procedure which finally suggests the applicability and quality of the model in terms of predicting the historical dataset and future response to the Model Inputs.

There is no complete set of numbers to define the Model Validation and different studies may use different sets of Model Validation metrics.

The basic set of Validation metrics is provided by Goodness of fit.

The important technique of Model Validation is Cross-Validation (also known as Blind Testing), which is widely used in scientific research and engineering practice.

This technique assumes that Source Dataset is split into two subsets: Training dataset and Validation dataset and perform the comparison of Goodness of fit over the two datasets.

It should be noted though that Source Dataset may not hold enough of representative events/occurrences to provide the opportunity for Cross-Validation and in this case the Goodness of fit over the Training dataset (which is the whole Source Dataset in this case) will be the only one available, thus increasing the risk of future Model Prediction.

See also

Human / Science / Formal Science / System Science  / System / Model

[ Goodness of fit ][ Cross-Validation ]

  • No labels