Experiment Data Processing by Introducing Integration Gray Regression Model and Its Application

The primary cause of test errors, in many instances, is relevant to the phenomenon that the quantity of experiment samples, especially in the engineering test with destructive effect, cannot meet the minimum required sample capacity ruled by the mathematical statistics. This study demonstrated an integration gray regression model (IGRM) and presented the modeling procedure, the IGRM is a combination model of main-rule model, which is based on the results of regression model, and system residual error model, which is based on the results of the deviation between truth values versus fitted values of the main-rule model. Both modeling calculation and laboratory experiments were performed to validate a good fitting effect of IGRM in the case of small sample size and poor information system. These results demonstrated that the fitting errors can be reduced to below 1% versus 10% of the regression modeling.