chapter 18

# chapter 18 - Chapter 18 Model Validation Important Article...

This preview shows pages 1–10. Sign up to view the full content.

Chapter 18 Model Validation

This preview has intentionally blurred sections. Sign up to view the full version.

View Full Document
Important Article McCarl, B.A. and J.D. Apland. "Validation of Linear Programming Models." Southern Journal of Agricultural Economics. 68,5(1986):155-164.
Background Model validation is important in any empirical analysis. Programming models frequently are superficially validated. However, validation is necessary for both predictive and prescriptive model use. Validation exercises almost always improve model performance and problem insight.

This preview has intentionally blurred sections. Sign up to view the full version.

View Full Document
Basic Model Maximize f(X) - g(Z) s.t. X -DY ≤ 0 GY - Z ≤ 0 AY ≤ b X,Y,Z ≥ 0 The optimal values, X*, Y*, Z are assumed to correspond to real world observations X’, Y’, Z’. The model also has shadow prices U*, V*, W* that correspond to real world values U’, V’, W’. X is some output measure. Z is some input measures.
Approaches to Validation Validation by Construction Validation by Results Validation Experiments

This preview has intentionally blurred sections. Sign up to view the full version.

View Full Document
General Approaches Validation approaches vary widely. The overall purpose is to test how well a model serves its intended purpose. For predictive models, validation tests can involve comparing model predictions to real world results. For prescriptive models, decision maker reliance is the ultimate validation test. Unfortunately, these tests can rarely be used because they are expensive and time-consuming (this is oftenthe reason for modeling in the first place). Thus, models are frequently validated using historical events. Although a model may have a broad range of potential uses, it may be valid only for a few of those uses. The validation process usually results in identification of valid applications.
Validation is Subjective Model validation is fundamentally subjective. Modelers choose the validity tests, the criteria for passing those tests, what model outputs to validate, what setting to test in, what data to use, etc. Thus, the assertion "the model was judged valid" can mean almost anything. Nonetheless, a model validation effort will reveal model strengths and weaknesses which is valuable to users and those who extract information from model results.

This preview has intentionally blurred sections. Sign up to view the full version.

View Full Document
Validation by Construction Validation by construct, as the sole method of validation, is justified by one of several assertions about modeling. The right procedures were used by the model builder. Usually this involves the assertion that the approach is consistent with industry, previous research and/or theory; and that the data were specified using reasonable scientific estimation or accounting procedures (deducing the model data from real world observations). Trial results indicate the model is behaving satisfactorily. This arises from a nominal examination of model results which indicates they do not contradict the modeler's, user's, and/or associated "experts" perceptions of reality.
Validation by Construction, Cont. Constraints were imposed which restrict the model to realistic

This preview has intentionally blurred sections. Sign up to view the full version.

View Full Document
This is the end of the preview. Sign up to access the rest of the document.

## This note was uploaded on 11/15/2011 for the course AGEC 7100 taught by Professor Duffy,p during the Fall '08 term at Auburn University.

### Page1 / 32

chapter 18 - Chapter 18 Model Validation Important Article...

This preview shows document pages 1 - 10. Sign up to view the full document.

View Full Document
Ask a homework question - tutors are online