TIME SERIES FORECASTING OF HIGHWAY ACCIDENT FATALITIES

USING TWELVE YEARS OF TIME SERIES DATA ON HIGHWAY FATALITIES, THE METHODOLOGY CURRENTLY EMPLOYED BY THE NATIONAL HIGHWAY TRAFFIC SAFETY ADMINISTRATIOON (NHTSA) TO FORECAST THE ANNUAL (CALENDAR YEAR) TOTAL OF HIGHWAY ACCIDENT FATALITIES WERE COMPARED WITH THOSE OBTAINED BY SEVERAL COMPUTER ROUTINES BASED ON EXPONENTIAL SMOOTHING TECHNIQUES AND AVAILABLE AT THE NATIONAL BUREAU OF STANDARDS. THE USE OF UNADJUSTED AND SEASONALLY ADJUSTED DATA WAS ALSO EXAMINED. IT I FOUND THAT THERE IS NO EVIDENCE TO LEAD TO ABANDONING THE PRESENT NHTSA METHODS IN FAVOR OF READILY AVAILABLE COMPUTER ROUTEINES BASED ON EXPONENTIAL SMOOTHING METHODS. OF THE METHODS EXAMINED IN STUDY, THE BEST RESULTS WERE OBTAINED WITH THE EXPSMOOTHING ROUTINE USING UNADJUSTED FATALITY DATA. /NTIS/

• Supplemental Notes:
  • Rept No NBSIR-73-138
• Corporate Authors:

  National Bureau of Standards /US

  /Inst for Applied Tech
  ,   United States 
• Authors:
  • Craw, A R
• Publication Date: 1973-3

Media Info

• Pagination: 39 p.

Subject/Index Terms

• TRT Terms: Computer programs; Exponential functions; Extrapolation; Fatalities; Forecasting; Methodology; Time series analysis; Traffic crashes
• Uncontrolled Terms: Motor vehicle accidents
• Subject Areas: Highways; Safety and Human Factors;

Filing Info

• Accession Number: 00224416
• Record Type: Publication
• Source Agency: National Technical Information Service
• Report/Paper Numbers: Final Rept
• Files: TRIS
• Created Date: Jul 31 1974 12:00AM