期货现货日数据，做var模型，lag interval默认是1-2，做出var模型后，lag length criteria 默认为8 ，结果如图所示，
| VAR Lag Order Selection Criteria | | | | |
| Endogenous variables: DLNF DLNS | | | | |
| Exogenous variables: C | | | | |
| Date: 01/08/16 Time: 14:18 | | | | |
| Sample: 1/02/2003 12/31/2015 | | | | |
| Included observations: 3129 | | | | |
| | | | | | | |
| | | | | | | |
| Lag | LogL | LR | FPE | AIC | SC | HQ |
| | | | | | | |
| | | | | | | |
| 0 | 14583.90 | NA | 3.07e-07 | -9.320483 | -9.316617 | -9.319095 |
| 1 | 15079.68 | 990.6139 | 2.24e-07 | -9.634821 | -9.623223 | -9.630658 |
| 2 | 15133.35 | 107.1639 | 2.17e-07 | -9.666568 | -9.647237* | -9.659629 |
| 3 | 15145.40 | 24.05497 | 2.16e-07 | -9.671716 | -9.644653 | -9.662002 |
| 4 | 15159.36 | 27.84170 | 2.15e-07 | -9.678083 | -9.643288 | -9.665594 |
| 5 | 15167.25 | 15.71532 | 2.14e-07 | -9.680566 | -9.638039 | -9.665302 |
| 6 | 15178.62 | 22.65150* | 2.13e-07* | -9.685279* | -9.635020 | -9.667240* |
| 7 | 15180.56 | 3.859591 | 2.13e-07 | -9.683962 | -9.625971 | -9.663147 |
| 8 | 15182.69 | 4.242157 | 2.14e-07 | -9.682768 | -9.617045 | -9.659178 |
| | | | | | | |
| | | | | | | |
当把lags to include 改大，*号分布的都很分散，不会集中在那个阶上，AIC最小的通常是11阶，请问要如何确认最佳的滞后阶数呢？研究了好久都没搞清楚，请各位大侠指教一下。
扫码加好友，拉您进群
全部版块
我的主页
收藏
