[时间序列分析][2]--趋势和(季节)因子
一句话概括今天要讲的内容:
对于有趋势的时间序列的数据,我们可以先把趋势给拟合出来,然后在对残差进行分析,看残差是否是稳定的,是否是白噪声序列。
我们对于趋势和因子举两个例子,分别来看一下:
一.趋势
下面这组数据是 “上海证券交易所1991年1月 - 2001 年10月每月末上证指数序列”
{130.44, 133.47, 120.19, 113.94, 114.83, 137.56, 143.8, 178.43, \
180.92, 218.6, 259.6, 292.75, 313.24, 364.66, 381.24, 445.38, \
1234.71, 1191.19, 1052.07, 823.27, 702.32, 507.25, 724.6, 780.39, \
1198.48, 1339.88, 925.91, 1358.78, 935.48, 1007.05, 881.07, 895.68, \
890.27, 814.82, 984.93, 833.8, 770.25, 770.98, 704.46, 592.56, \
556.26, 469.29, 333.92, 785.33, 791.15, 654.98, 683.59, 647.87, \
562.59, 549.26, 646.92, 579.93, 700.51, 630.58, 695.55, 723.87, \
722.43, 717.32, 641.13, 555.29, 537.34, 552.93, 556.39, 681.16, \
643.65, 804.25, 822.48, 809.94, 875.52, 976.71, 1032.95, 917.02, \
964.74, 1040.27, 1234.62, 1393.75, 1285.18, 1250.27, 1189.76, \
1221.06, 1097.38, 1180.39, 1139.63, 1194.1, 1222.91, 1206.53, \
1243.01, 1343.44, 1411.2, 1339.2, 1316.91, 1150.22, 1242.09, 1217.31, \
1247.42, 1146.7, 1134.67, 1090.09, 1158.05, 1120.92, 1279.32, \
1689.42, 1601.45, 1627.12, 1570.7, 1504.56, 1434.97, 1366.