Join/Renew
Membership
Subscribe to
Mailing List
| Resources >> Time Series Data • M-Competition • M2-Competition • M3-Competition |
M2-CompetitionThe M2-Competition called for participants to produce forecasts for quarterly and monthly series from actual companies. Judgmental methods were added to the complement of extrapolation methods examines in earlier studies and two rounds of forecasts were produced.File M2C87.DAT contains the first set of data 6 monthly series with 70 data points ending 1987/October 6 quarterly series with 163 data points ending 1987/III 6 monthly series with 82 data points ending 1987/October 1 monthly series with 33 data points ending 1987/October 6 monthly series with 70 data points ending 1987/October 2 monthly series with 82 data points ending 1987/October 2 monthly series with 225 data points ending 1987/September on which the following forecasts should be done : 14 forecasting horizons for the monthly series (ending 1987/OCT) 15 forecasting horizons for the 2 last monthly series (ending 1987/SEP) 5 forecasting horizons for the quarterly series (ending 1987/III) File M2C88.DAT contains the second set of data : 6 monthly series with 82 data points ending 1988/October 6 quarterly series with 167 data points ending 1988/III 6 monthly series with 94 data points ending 1988/October 1 monthly series with 45 data points ending 1988/October 6 monthly series with 82 data points ending 1988/October 2 monthly series with 94 data points ending 1988/October 2 monthly series with 237 data points ending 1988/September on which the following forecasts should be done : 14 forecasting horizons for the monthly series (ending 1988/OCT) 15 forecasting horizons for the 2 last monthly series (ending 1988/SEP) 5 forecasting horizons for the quarterly series (ending 1988/III) Example of the format the series are written: PANTER 12 70 3619.800 7466.700 10579.000 7369.100 14987.700 8398.800 7254.300 8292.600 13658.000 11680.200 16948.100 14821.000 12240.700 10781.500 14179.000 15539.500 16254.300 20356.801 17812.301 21209.900 20711.100 22659.301 25397.500 25854.301 27785.199 30090.100 39672.801 37833.301 47240.699 36976.500 35719.801 27016.000 23725.900 35946.898 28540.699 27216.000 32648.100 31706.199 35259.301 37001.199 45906.199 41697.500 39081.500 42584.000 47063.000 31780.199 33325.898 30859.301 45029.602 39048.102 38428.398 45209.898 52727.199 48446.898 36803.699 38739.500 48317.301 33040.699 34324.699 44755.602 26744.400 34693.801 44292.602 47581.500 34780.199 31848.100 22879.000 24228.400 24228.400 29623.500 1982/JAN to 1987/OCT AUTOMOBILE INDUSTRY : CAR MODEL Code to read each series in FORTRAN: DIMENSION X(300) CHARACTER*9 FILE0 CHARACTER*60 TEXT1,TEXT2 C C READ THE FIRST LINE OF DATA IDENTIFYING EACH SERIES AND ITS CHARACTERISTICS C 10 READ(20,100) FILE0,LSEAS,N 100 FORMAT(1XA9,2I5) C C READ THE N ORIGINAL DATA USED TO DEVELOP THE FORECASTING MODEL C READ(20,200) (X(I),I=1,N) C C READ 2 LINES OF TEXT C READ(20,300) TEXT1,TEXT2 200 FORMAT (1X,6F12.6) 300 FORMAT (1X,A60) END Where: FILE0 : Name of the series LSEAS : Length of seasonality ( 4 for Quarterly, 12 for Monthly) N : Number of observations X(1) to X(N) : N Original data TEXT1,TEXT2 : 2 lines of text of 60 characters which gives the beginning and ending dates of the data and explains their origin. |
