SMOOTHIE: Interactive Time--Series Analysis



by Richard S. Barr and James Collins

SMOOTHIE is an interactive program for time--series data that permits the application of several forecasting models to a set of data. The user can apply single and double exponential smoothing, moving average, trend, arithmetic mean, and complete decomposition methods to a given data series.



Running SMOOTHIE


Prepare a data file using a text editor and the instructions below. To execute the program, type the following at the DOS prompt:

SMOOTHIE filename

where filename is the name of the data file. If filename is omitted, operating instructions are displayed.



Data File Organization


The data file contains a set of time--series data in free format. That is, it consists of a series of observed values for a single variable, taken at uniform time intervals, given in chronological order, and separated by blanks or on separate lines.

The following example file will be used below to illustrate the use of the program, and is assumed to be in the file named MYFILE.

13.1 12 15.8 11.4 14 15.3 17.6 13 15.2 16.4 19.2 14.7  15.6 15.5 17.2 16.7 16.8 16.5 17.4 14.4  17.0 19.4 19.2 16.7

There are 25 observations of a variable, given in time--series order.



Program Operation and Report Interpretation


In the sections that follow, user input will be shown in underlined boldface. To process the above data file, we enter:

SMOOTHIE myfile

The initial screen displays:

       *** SMTHIE: INTERACTIVE TIME--SERIES FORECASTING ***

                by RichOOard S. Barr and Jim Collins

Processing data file  smoothie.dat

24 observations were read

How many observations to use in the mean error calculation  (enter 0 for the maximum possible)? 5

The user is asked to enter the number of observations to be used in calculating cumulative error values, such as mean squared error (MSE). Since the models result in varying numbers of historical error terms, one may wish to use a subset in computing cumulatives, if the values are to be compared across models. We have asked that only the most recent five error values be used.

       

 *** SMOOTHIE: INTERACTIVE TIME--SERIES FORECASTING ***

The following forecasting models are available:

Moving average

Exponential smoothing (single)                     

Smoothing with trend (double)                      

Trend                                              

Arithmetic mean                                    

Decomposition                         

           

Select model by typing the first letter, or ESC to quit:              

Any of the six models listed may be applied to the dataset. A model is selected by typing its first letter, such as “M” for moving average.

Each model type results in a different forecast methodology and set of reports. The output for each model as applied to our dataset is given below.

NOTE:

To prevent reports from scrolling off of the screen before viewing by the user, output pauses at the end of each full screen. This condition is indicated by a “&” symbol, and the user may press any key to continue.

M, Moving Average Forecasts


Moving average forecasting is selected at the main menu by the “M” key. The user is then asked to enter the number of periods to be used in the averaging process. In our example, we will perform a 3--period moving average.

           *** MOVING AVERAGE METHOD ***

Number of periods for your moving average: 3

              *** SUMMARY REPORT ***

Period         Actual         Forecast          Error

Number        Observn         for Next          (Forecast--

                             Period            Actual)

------ ------- -------- ---------

  1           13.100

  2           12.000

  3           15.800          13.633

  4           11.400          13.067            -2.233

  5           14.000          13.733             0.933

  6           15.300          13.567             1.567

  7           17.600          15.633             4.033

  8           13.000          15.300            -2.633

  9           15.200          15.267            -0.100

 10           16.400          14.867             1.133

 11           19.200          16.933             4.333

 12           14.700          16.767            -2.233

 13           15.600          16.500            -1.167

 14           15.500          15.267            -1.000

 15           17.200          16.100             1.933

 16           16.700          16.467             0.600

 17           16.800          16.900             0.333

 18           16.500          16.667            -0.400

 19           17.400          16.900             0.733

 20           14.400          16.100            -2.500

 21           17.000          16.267             0.900

 22           19.400          16.933             3.133

 23           19.200          18.533             2.267

 24           16.700          18.433            -1.833

             *** FORECAST ERROR MEASURES ***

                             Specified        Maximum

Number of periods                     5             21

Mean squared error (MSE)          5.075          4.249

Mean absolute error (MAD)          2.127          1.714

Mean absolute pct error (MAPE)    12.318         10.666

Mean error/Forecast bias          0.393          0.371

E, Single Exponential Smoothing


When “E" is chosen from the main menu, and a smoothing constant (a) of 0.5 is selected, the following reports result.

            *** EXPONENTIAL SMOOTHING (SINGLE) ***

What smoothing constant do you wish to use (0 < a < 1)? 0.5

                 *** SUMMARY REPORT ***

Period         Actual         Forecast          Error

Number        Observn         for Next          (Forecast-

------ ------- -------- ---------

  1           13.100          13.100

  2           12.000          12.550            -1.100

  3           15.800          14.175             3.250

  4           11.400          12.788            -2.775

  5           14.000          13.394             1.212

  6           15.300          14.347             1.906

  7           17.600          15.973             3.253

  8           13.000          14.487            -2.973

  9           15.200          14.843             0.713

 10           16.400          15.622             1.557

 11           19.200          17.411             3.578

 12           14.700          16.055            -2.711

 13           15.600          15.828            -0.455

 14           15.500          15.664            -0.328

 15           17.200          16.432             1.536

 16           16.700          16.566             0.268

 17           16.800          16.683             0.234

 18           16.500          16.591            -0.183

 19           17.400          16.996             0.809

 20           14.400          15.698            -2.596

 21           17.000          16.349             1.302

 22           19.400          17.874             3.051

 23           19.200          18.537             1.326

 24           16.700          17.619            -1.837

             *** FORECAST ERROR MEASURES ***

                             Specified        Maximum

Number of periods                     5             23

Mean squared error (MSE)          4.575          4.063

Mean absolute error (MAD)          2.022          1.694

Mean absolute pct error (MAPE)     11.864         10.863

Mean error/Forecast bias          0.249          0.393

S, Smoothing with Trend (Double)


Double exponential smoothing uses both a smoothing constant, a, but a trend emphasis constant, b. Both values are between 0 and 1.

        *** EXPONENTIAL SMOOTHING WITH TREND (DOUBLE) ***

What DATA smoothing constant do you wish to use (0 < a < 1)? 0.5

What TREND smoothing constant do you wish to use (0 < b < 1)? 0.6

Forecast using 18.340528 + 0.094073 * (number of periods beyond 24)

               *** SUMMARY REPORT ***

Period         Actual         Forecast          Error

Number        Observn         for Next          (Forecast-

                             Period            Actual)

------ ------- -------- ---------

  1           13.100          13.100

  2           12.000          12.220            -1.100

  3           15.800          14.754             3.580

  4           11.400          12.815            -3.354

  5           14.000          13.501             1.185

  6           15.300          15.034             1.799

  7           17.600          17.720             2.566

  8           13.000          15.347            -4.720

  9           15.200          15.217            -0.147

 10           16.400          16.106             1.183

 11           19.200          18.879             3.094

 12           14.700          16.762            -4.179

 13           15.600          15.805            -1.162

 14           15.500          15.185            -0.305

 15           17.200          16.329             2.015

 16           16.700          16.763             0.371

 17           16.800          17.041             0.037

 18           16.500          16.867            -0.541

 19           17.400          17.391             0.533

 20           14.400          15.255            -2.991

 21           17.000          16.011             1.745

 22           19.400          18.605             3.389

 23           19.200          19.981             0.595

 24           16.700          18.435            -3.281

             *** FORECAST ERROR MEASURES ***

                             Specified        Maximum

Number of periods                     5             23

Mean squared error (MSE)          6.919          5.568

Mean absolute error (MAD)          2.400          1.907

Mean absolute pct error (MAPE)     14.250         12.445

Mean error/Forecast bias         -0.109          0.014

T, Trend


Forecasting with a trend line only involved fitting a simple regression line to the observed points, using the period number as the independent variable.

           *** LEAST--SQUARES TREND LINE ***

Trend equation:  T = 13.352174 + 0.198826 * t

Sample coefficient of determination (R--squared) = 0.433809

               *** SUMMARY REPORT ***

Period         Actual         Forecast          Error

Number        Observn         for Next          (Forecast-

                             Period            Actual)

------ ------- -------- ---------

  1           13.100          13.100

  2           12.000          10.900            -1.100

  3           15.800          16.333             4.900

  4           11.400          12.750            -4.933

  5           14.000          13.620             1.250

  6           15.300          14.860             1.680

  7           17.600          16.786             2.740

  8           13.000          15.546            -3.786

  9           15.200          15.731            -0.346

 10           16.400          16.313             0.669

 11           19.200          17.715             2.887

 12           14.700          17.192            -3.015

 13           15.600          17.069            -1.592

 14           15.500          16.935            -1.569

 15           17.200          17.275             0.265

 16           16.700          17.408            -0.575

 17           16.800          17.528            -0.608

 18           16.500          17.551            -1.028

 19           17.400          17.753            -0.151

 20           14.400          17.313            -3.353

 21           17.000          17.437            -0.313

 22           19.400          17.973             1.963

 23           19.200          18.389             1.227

 24           16.700          18.323            -1.689

             *** FORECAST ERROR MEASURES ***

                             Specified        Maximum

Number of periods                     5             23

Mean squared error (MSE)          3.910          5.225

Mean absolute error (MAD)          1.709          1.810

Mean absolute pct error (MAPE)   10.350         12.075

Mean error/Forecast bias         -0.433         -0.282

A, Arithmetic Mean


The average of all previous observations can be used as a forecast, as follows.                                    

*** ARITHMETIC MEAN AS A FORECAST ***

The overall mean of Y is 15.837500

               *** SUMMARY REPORT ***

Period         Actual         Forecast          Error

Number        Observn         for Next          (Forecast-

                             Period            Actual)

------ ------- -------- ---------   1           13.100          13.100

  2           12.000          12.550            -1.100

  3           15.800          13.633             3.250

  4           11.400          13.075            -2.233

  5           14.000          13.260             0.925

  6           15.300          13.600             2.040

  7           17.600          14.171             4.000

  8           13.000          14.025            -1.171

  9           15.200          14.156             1.175

 10           16.400          14.380             2.244

 11           19.200          14.818             4.820

 12           14.700          14.808            -0.118

 13           15.600          14.869             0.792

 14           15.500          14.914             0.631

 15           17.200          15.067             2.286

 16           16.700          15.169             1.633

 17           16.800          15.265             1.631

 18           16.500          15.333             1.235

 19           17.400          15.442             2.067

 20           14.400          15.390            -1.042

 21           17.000          15.467             1.610

 22           19.400          15.645             3.933

 23           19.200          15.800             3.555

 24           16.700          15.837             0.900

             *** FORECAST ERROR MEASURES ***

                             Specified        Maximum

Number of periods                     5             23

Mean squared error (MSE)          6.519          5.164

Mean absolute error (MAD)          2.208          1.930

Mean absolute pct error (MAPE      12.177         11.761

Mean error/Forecast bias          1.791          1.437

                                   

D, Decomposition                                    


Time--series decomposition separates the observations into four components: trend, seasonal, cyclical, and irregular/random. The multiplicative model is of the form: . The user can then assemble a forecast for a future period by estimating its cyclical index (a value of 1.0 can be used to ignore this component). A detailed historical decomposition report is available as a use option.

For computing the seasonal component, the user must indicate whether the observations are monthly or quarterly data values. The sample dataset is based on quarterly observations.

*** TIME SERIES DECOMPOSITION ***

Is the data monthly or quarterly (M/Q)? Q

Quarterly Trend equation:  T = 13.657547 + 0.178496 * Period Sample coefficient of determination (R--squared) = 0.826790

Season       Seasonal Index

--—--—--—       --—--—--—--—--—--—--—

Quarter 1          0.9794

Quarter 2          1.0167

Quarter 3          1.1347

Quarter 4          0.8693

Modified means used in 4 seasonal index calculations

Mean irregular index = 1.005263

Do you want a detailed report (Y/N)? Y

                D E C O M P O S I T I O N    S U M M A R Y

Period  Actual  Trend   Seasonal   Cyclical   Forecast      Error

3    15.800   14.193      1.135      0.929     16.105     -0.305

4    11.400   14.372      0.869      0.954     12.493     -1.093

-- ------ ------ ----- ----- ------ ------

5    14.000   14.550      0.979      0.986     14.250     -0.250

6    15.300   14.729      1.017      1.003     14.974      0.326

7    17.600   14.907      1.135      1.015     16.915      0.685

-- ------ ------ ----- ----- ------ ------

9    15.200   15.264      0.979      1.032     14.949      0.251

10    16.400   15.443      1.017      1.047     15.700      0.700

11    19.200   15.621      1.135      1.051     17.725      1.475

12    14.700   15.800      0.869      1.036     13.734      0.966

-- ------ ------ ----- ----- ------ ------

13    15.600   15.978      0.979      1.001     15.648     -0.048

14    15.500   16.156      1.017      0.990     16.426     -0.926

15    17.200   16.335      1.135      1.004     18.535     -1.335

16    16.700   16.513      0.869      1.010     14.355      2.345

-- ------ ------ ----- ----- ------ ------

17    16.800     16.692      0.979      1.008     16.348    0.452

18    16.500     16.870      1.017      0.982     17.152   -0.652

19    17.400     17.049      1.135      0.956     19.345   -1.945

20    14.400     17.227      0.869      0.969     14.975   -0.575

-- ------ ------ ----- ----- ------ ------

21    17.000     17.406      0.979      0.992     17.047   -0.047

22    19.400     17.584      1.017      1.012     17.878    1.522

Forecast based on trend and seasonal only

Mean squared error for periods 3 through 22 = 1.040293

           *** INTERACTIVE FORECASTING ***

Period to forecast or 0? 30

Cyclical index to use or 1 to ignore? 1   

Forecast =       Trend * Seasonal Index * Cyclical Index     19.329 =      19.012 *        1.017   *       1.000

Period to forecast or 0? 0

Exiting SMOOTHIE


To exit the system, press the Escape key from the main menu.