Table A-1: Time Series Characteristics of TSI Data Components
Thursday, September 6, 2012
Table A-1: Time Series Characteristics of TSI Data Components
| Mode | ARIMA model | Moving seasonality? | MCD1 | Model2: additive or multiplicative? | Trading days (TD) & Holidays (H)3 |
|---|---|---|---|---|---|
| Rail passenger | (011) (011) | No | 5 | M | No TD; all H |
| Rail freight | |||||
| Carloads | No valid model yet | No | 5 | A | TD, No H |
| Intermodals | (011) (011) | No | 3 | A | No TD; Easter & Thanksgiving |
| Trucking | (011) (011) | No | 3 | A | TD; Easter |
| Waterborne | (011) (011) | No | 12 | A | TD; No H |
| Transit | (011) (011) | No | 4 | A | TD; No H |
| Aviation | |||||
| Freight | (011) (011) | No | 3 | M | TD and Easter & Thanksgiving |
| Passenger | (011) (011) | Yes, at 1% level | 4 | M | No TD, All 3 H |
| Pipeline | |||||
| Natural gas | (012) (011) | No | 7 | M | TD; No H |
| Petroleum | (011) (011) | No | 5 | M | TD, No H |
SOURCE: U.S. Department of Transportation, Research and Innovative Technology Administration, Bureau of Transportation Statistics, http://www.bts.gov.
1 MCD=Months to cyclical dominance. 3 months is normal for the trend to equal and surpass the irregularity; higher than 3 shows irregularity too high.
2 M=Multiplicative Model A=Additive Model
3 TD=Trading Days, H=(Moving) Holidays: Easter, Labor Day, and Thanksgiving. Beginning January 27, 2005, Thanksgiving uses -3 for value, Easter is set at 2 for Trucking; otherwise, equals 7; Labor Day is always set at 4.