Analysis Summary Data variable: AUTOADJ/CPI Number of observations = 314 Start index = 1/70 Sampling interval = 1.0 month(s) Length of seasonality = 12 Forecast Summary ---------------- Forecast model selected: Constant mean = 1.40164 + 1 regressor Number of forecasts generated: 0 Number of periods withheld for validation: 26 Estimation Validation Statistic Period Period -------------------------------------------- MSE 4.10166 10.6607 MAE 1.64644 3.05438 MAPE 8.3251 9.9606 ME -1.43836E-14 3.05438 MPE -1.041 9.9606 Trend Model Summary Parameter Estimate Stnd. Error t P-value ---------------------------------------------------------------------------- Constant 1.40164 0.709646 1.97513 0.049215 INCOME/CPI 0.0782192 0.0028257 27.6814 0.000000 ---------------------------------------------------------------------------- The StatAdvisor --------------- This procedure will forecast future values of AUTOADJ/CPI. The data cover 314 time periods. Currently, a mean model has been selected. This model assumes that the best forecast for future data is given by the average of all previous data. You can select a different forecasting model by pressing the alternate mouse button and selecting Analysis Options. The output summarizes the statistical significance of the terms in the forecasting model. Terms with P-values less than 0.05 are statistically significantly different from zero at the 95% confidence level. In this case, the P-value for the mean is less than 0.05, so it is significantly different from 0.0. The model also includes one independent regression variable. Since INCOME/CPI has a P-value which is less than 0.05, it is statistically significant at the 95% confidence level. The table also summarizes the performance of the currently selected model in fitting the previous data. It displays: (1) the mean squared error (MSE) (2) the mean absolute error (MAE) (3) the mean absolute percentage error (MAPE) (4) the mean error (ME) (5) the mean percentage error (MPE) Each of the statistics is based on the one-ahead forecast errors, which are the differences between the data value at time t and the forecast of that value made at time t-1. The first three statistics measure the magnitude of the errors. A better model will give a smaller value. The last two statistics measure bias. A better model will give a value close to 0.0. In this case, the model was estimated from the first 288 data values. 26 data values at the end of the time series were withheld to validate the model. The table shows the error statistics for both the estimation and validation periods. If the results are considerably worse in the validation period, it means that the model is not likely to perform as well as otherwise expected in forecasting the future.