Discriminating between second-order model with variation in model parameters based on central tendency estimation

Study of discriminating between second-order model with or without interaction for central tendency estimation was presented using ordinary least square for estimation of the model parameters. The research considered two different sets of data small sample size which is data of unemployment rate as response, inflation rate and exchange rate as the predictors from 2007 to 2018 and large sample size which is data of flow-rate on hydrate formation for Niger Delta deep offshore field. The R^2, AIC, SBC and SSE were applied for both data sets to test for the adequacy of the models. The small data was used as illustration 1 and the large data was used as illustration 2. It was revealed that, model centered on mean with interaction proved better than mean model without interaction. Median and mode values were found to be equal, as a result, the estimates of the median models were equal to the estimates of the mode models in all cases for both large and small data. The models centered on median and mode with interaction were better than those without interaction for both illustrations. Mean and mode models with or without interactions were found better than the mean models with or without interactions for both illustrations. The joint effect of inflation and exchange rate were found to be insignificant to the unemployment rate in Nigeria, while that of the interaction in median and mode models were seen better than that of mean model with a percentage difference of 57.1 for models with interaction. The intercept for mean model with interaction was found less than that of median and mode models with a percentage difference of 61 approximately. The intercept for median and mode model without interaction are better than mean model with a percentage difference of 55.8.