This is the same as Example 1, except now class autoARIMA
uses Method 2 with a possible seasonal adjustment. As a result, the unadjusted model with p = 3, q = 2, s = 1, d = 0
is chosen as optimum.
import java.util.*; import com.imsl.stat.*; public class AutoARIMAEx2 { public static void main(String args[]) throws Exception { int nOutliers; double aic, RSE, constant; int[] optimumModel; int[][] outlierStatistics; double[] outlierForecast, ar, ma; double[] psiWeights, probabilityLimits; int[] arOrders = {0, 1, 2, 3}; int[] maOrders = {0, 1, 2, 3}; int[] periods = {1, 2}; int[] orders = {0, 1, 2}; double[] x = { 12.8, 12.2, 11.9, 10.9, 10.6, 11.3, 11.1, 10.4, 10.0, 9.7, 9.7, 9.7, 11.1, 10.5, 10.3, 9.8, 9.8, 10.4, 10.4, 10.0, 9.7, 9.3, 9.6, 9.7, 10.8, 10.7, 10.3, 9.7, 9.5, 10.0, 10.0, 9.3, 9.0, 8.8, 8.9, 9.2, 10.4, 10.0, 9.6, 9.0, 8.5, 9.2, 9.0, 8.6, 8.3, 7.9, 8.0, 8.2, 9.3, 8.9, 8.9, 7.7, 7.6, 8.4, 8.5, 7.8, 7.6, 7.3, 7.2, 7.3, 8.5, 8.2, 7.9, 7.4, 7.1, 7.9, 7.7, 7.2, 7.0, 6.7, 6.8, 6.9, 7.8, 7.6, 7.4, 6.6, 6.8, 7.2, 7.2, 7.0, 6.6, 6.3, 6.8, 6.7, 8.1, 7.9, 7.6, 7.1, 7.2, 8.2, 8.1, 8.1, 8.2, 8.7, 9.0, 9.3, 10.5, 10.1, 9.9, 9.4, 9.2, 9.8, 9.9, 9.5, 9.0, 9.0, 9.4, 9.6, 11.0, 10.8, 10.4, 9.8, 9.7, 10.6, 10.5, 10.0, 9.8, 9.5, 9.7, 9.6, 10.9, 10.3, 10.4, 9.3, 9.3, 9.8, 9.8, 9.3, 8.9, 9.1, 9.1, 9.1, 10.2, 9.9, 9.4}; double[] exactForecast = {8.7, 8.6, 9.3, 9.1, 8.8, 8.5}; int[] times = { 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 57, 58, 59, 60, 61, 62, 63, 64, 65, 66, 67, 68, 69, 70, 71, 72, 73, 74, 75, 76, 77, 78, 79, 80, 81, 82, 83, 84, 85, 86, 87, 88, 89, 90, 91, 92, 93, 94, 95, 96, 97, 98, 99, 100, 101, 102, 103, 104, 105, 106, 107, 108, 109, 110, 111, 112, 113, 114, 115, 116, 117, 118, 119, 120, 121, 122, 123, 124, 125, 126, 127, 128, 129, 130, 131, 132, 133, 134, 135}; AutoARIMA autoArima = new AutoARIMA(times, x); autoArima.setCriticalValue(3.8); autoArima.setMaximumARLag(5); autoArima.setPeriods(periods); autoArima.setDifferenceOrders(orders); autoArima.compute(arOrders, maOrders); autoArima.forecast(6); nOutliers = autoArima.getNumberOfOutliers(); aic = autoArima.getAIC(); optimumModel = autoArima.getOptimumModelOrder(); outlierStatistics = autoArima.getOutlierStatistics(); RSE = autoArima.getResidualStandardError(); outlierForecast = autoArima.getForecast(); psiWeights = autoArima.getPsiWeights(); probabilityLimits = autoArima.getDeviations(); constant = autoArima.getConstant(); ar = autoArima.getAR(); ma = autoArima.getMA(); System.out.printf("%nMethod 2: Grid search with "+ "differencing%n"); System.out.printf("%nOptimum Model: p=%d, q=%d, s=%d, d=%d%n", optimumModel[0], optimumModel[1], optimumModel[2], optimumModel[3]); System.out.printf("%nNumber of outliers:%3d%n%n", nOutliers); System.out.printf("Outlier statistics:%n"); System.out.printf(" Time%4sType%n", " "); for (int i=0; i<nOutliers; i++) System.out.printf("%5d%8d%n", outlierStatistics[i][0], outlierStatistics[i][1]); System.out.printf(Locale.ENGLISH, "%nAIC:%12.6f%n", aic); System.out.printf(Locale.ENGLISH, "RSE:%12.6f%n%n", RSE); System.out.printf("%5sParameters%n", " "); System.out.printf(Locale.ENGLISH, " constant:%10.6f%n", constant); for (int i=0; i<ar.length; i++) System.out.printf(" ar[%d]:%13.6f%n", i, ar[i]); for (int i=0; i<ma.length; i++) System.out.printf(" ma[%d]:%13.6f%n", i, ma[i]); System.out.printf("%n%n%6s* * * Forecast Table * * *%n", " "); System.out.printf("%2sExact%3sforecast%5slimits%8spsi%n", " ", " ", " ", " "); for (int i=0; i<outlierForecast.length; i++) System.out.printf(Locale.ENGLISH, "%7.4f%11.4f%11.4f%11.4f%n", exactForecast[i], outlierForecast[i], probabilityLimits[i], psiWeights[i]); } }
Method 2: Grid search with differencing Optimum Model: p=3, q=2, s=1, d=0 Number of outliers: 1 Outlier statistics: Time Type 109 0 AIC: 408.108176 RSE: 0.412456 Parameters constant: 0.554459 ar[0]: 1.940615 ar[1]: -1.898025 ar[2]: 0.897791 ma[0]: 1.115803 ma[1]: -0.911902 * * * Forecast Table * * * Exact forecast limits psi 8.7000 9.1085 0.8084 0.8248 8.6000 9.1715 1.0479 0.6145 9.3000 9.5039 1.1597 0.5248 9.1000 9.7677 1.2349 0.5926 8.8000 9.7051 1.3245 0.7056 8.5000 9.3817 1.4421 0.7157Link to Java source.