Normalvy
MARC-vy
Time series analysis and forecast of annual crash fatalities Liu, Cejun ; Chen, Chou-Lin
Utgivningsinformation: Washington DC National Center for Statistics and Analysis, NCSA, 2004; Traffic Safety Facts. Research note March 2004, Beskrivning: 4 sÄmnen: Onlineresurser: Abstrakt: This research note uses two Time Series techniques, Holt-Winters (HW) Algorithm and Autoregressive Moving Average Model (ARMA), to predict annual motor vehicle crash fatalities. Based on the monthly Fatality Analysis Reporting System (FARS) data from 1975 to 2001, the estimated fatalities are 42,675 and 42,876 respectively in 2002. These estimates are very close to the true counts, as compared to the 2002 fatalities of 42,815. Incorporating the actual 2002 fatalities in the data series, the forecast values in 2003, 41,349 and 41,876, show a decline from the fatalities of 2002.Inga fysiska exemplar för denna post
This research note uses two Time Series techniques, Holt-Winters (HW) Algorithm and Autoregressive Moving Average Model (ARMA), to predict annual motor vehicle crash fatalities. Based on the monthly Fatality Analysis Reporting System (FARS) data from 1975 to 2001, the estimated fatalities are 42,675 and 42,876 respectively in 2002. These estimates are very close to the true counts, as compared to the 2002 fatalities of 42,815. Incorporating the actual 2002 fatalities in the data series, the forecast values in 2003, 41,349 and 41,876, show a decline from the fatalities of 2002.