forecastmethods_techniques
TRANSCRIPT
-
7/25/2019 ForecastMETHODS_TECHNIQUES
1/116
All Rights Reserved, Indian
-
7/25/2019 ForecastMETHODS_TECHNIQUES
2/116
All Rights Reserved, Indian
Those who have knowledge dont predict.
Those who predict dont have knowledge.
-
7/25/2019 ForecastMETHODS_TECHNIQUES
3/116
All Rights Reserved, Indian
I think there is a world market for may be 5 computer
- Thomas Watson, Chairman of IBM 19
Computers in future weigh no more than 1.5 tons
- Popular Mechanics, 1949
640K ought to be enough for everybody
- Bill Gates, 1981???
-
7/25/2019 ForecastMETHODS_TECHNIQUES
4/116
All Rights Reserved, Indian
Forecasting
Forecasting is a process of estimation of an unknown evensuch as demand for a product.
Forecasting is commonly used to refer time series data.
Time series is a sequence of data points measured at sucintervals.
-
7/25/2019 ForecastMETHODS_TECHNIQUES
5/116 All Rights Reserved, Indian
Corporate
Strategy
Product and Market
Planning
Business
Forecasting
Product and Market Strate
Aggregate Production
Planning
Master ProductionPlanning
Materials
Requirement
Planning
Aggregate
Forecasting
ItemForecasting
Demand
forecasting at SKU
Level
Resource Planning -
Medium to Long Range
Planning
Production Capacity - ShoRange Planning
Capacity Requirement
Planning
-
7/25/2019 ForecastMETHODS_TECHNIQUES
6/116 All Rights Reserved, Indian
Forecasting methods
Qualitative Techniques.
Expert opinion (or Astrologers)
Quantitative Techniques.
Time series techniques such as exponential smoothing
Casual Models.
Uses information about relationship between system elements
(e.g regression).
-
7/25/2019 ForecastMETHODS_TECHNIQUES
7/116 All Rights Reserved, Indian
-
7/25/2019 ForecastMETHODS_TECHNIQUES
8/116 All Rights Reserved, Indian
Why Time Series Analysis ?
Time series analysis helps to identify and explain:
Any systemic variation in the series of data which is due to
Cyclical pattern that repeat.
Trends in the data. Growth rates in the trends.
-
7/25/2019 ForecastMETHODS_TECHNIQUES
9/116 All Rights Reserved, Indian
Time Series Components
Trend Cyclical
Seasonal Irregular
-
7/25/2019 ForecastMETHODS_TECHNIQUES
10/116 All Rights Reserved, Indian
Trend Component
Persistent upward or downward pattern
Due to consumer behaviour, population, economy, techno
-
7/25/2019 ForecastMETHODS_TECHNIQUES
11/116
All Rights Reserved, Indian
Cyclical Component
Repeating up and down movements.
Due to interaction of factors influencing economy such
Usually 2-10 years duration.
-
7/25/2019 ForecastMETHODS_TECHNIQUES
12/116
All Rights Reserved, Indian
Seasonal Component
Regular pattern of up and down movements.
Due to weather, customs, festivals etc.
Occurs within one year.
-
7/25/2019 ForecastMETHODS_TECHNIQUES
13/116
All Rights Reserved, Indian
Seasonal Vs Cyclical
When a cyclical pattern in the data has a period of lesyear, it is referred as seasonal variation.
When the cyclical pattern has a period of more than orefer to it as cyclical variation.
-
7/25/2019 ForecastMETHODS_TECHNIQUES
14/116
All Rights Reserved, Indian
Irregular Component
Erratic fluctuations
Due to random variation or unforeseen events
White Noise
-
7/25/2019 ForecastMETHODS_TECHNIQUES
15/116
All Rights Reserved, Indian
Time(a) Trend
Time(d) Trend with seasonal pattern
Time(c) Seasonal pattern
Time(b) Cycle
Demand
Demand
Dem
and
Demand
Randommovement
More
than
one
year gap
Less
than
one
year gap
-
7/25/2019 ForecastMETHODS_TECHNIQUES
16/116
All Rights Reserved, Indian
-
7/25/2019 ForecastMETHODS_TECHNIQUES
17/116
All Rights Reserved, Indian
Time Series DataAdditive and Multiplicati
Random
t
Cyclical
t
ySeasonalit
t
Trend
t RCST tY
ModelgForecastinAdditive1.
Random
t
Cyclical
t
ySeasonalit
t
Trend
t RCST tY
ModelgForecastintiveMultiplica2.
-
7/25/2019 ForecastMETHODS_TECHNIQUES
18/116
All Rights Reserved, Indian
Time Series Data Decomposition
ySeasonalit
t
Trend
t ST tY
ModelgForecastintiveMultiplica
t
tt
T
YS
Yt / Tt is called
deseasonalized data
-
7/25/2019 ForecastMETHODS_TECHNIQUES
19/116
All Rights Reserved, Indian
Time Series Techniques
Moving Average.
Exponential Smoothing.
Auto-regression Models (AR Models).
ARIMA (Auto-regressive Integrated Moving Averag
-
7/25/2019 ForecastMETHODS_TECHNIQUES
20/116
All Rights Reserved, Indian
Moving Average Method
-
7/25/2019 ForecastMETHODS_TECHNIQUES
21/116
All Rights Reserved, Indian
Moving Average (Rolling Average)
Simple moving average. Used mainly to capture trend and smooth short term
Most recent data are given equal weights.
Weighted moving average Uses unequal weights for data
-
7/25/2019 ForecastMETHODS_TECHNIQUES
22/116
All Rights Reserved, Indian
Data
Demand for continental breakfast at the Die Another Da
Daily data between 1 October 201423 January 2015 (
-
7/25/2019 ForecastMETHODS_TECHNIQUES
23/116
All Rights Reserved, Indian
Simple moving average
iperiodtimetoingcorrespondDataY
1tperiodforForecastF
Yn
1F
i
1t
t
1nti
i1t
The forecast for period t+1 (Ft+1
) is given by the average of
recent data.
-
7/25/2019 ForecastMETHODS_TECHNIQUES
24/116
All Rights Reserved, Indian
Simple moving average
iperiodtimetoingcorrespondDataY
1tperiodforForecastF
Yn
1F
i
1t
t
1nti
i1t
The forecast for period t+1 (Ft+1
) is given by the average of
recent data.
-
7/25/2019 ForecastMETHODS_TECHNIQUES
25/116
All Rights Reserved, Indian
Demand for Continental Breakfast at DAD Hosp
Moving Average
0
10
20
30
40
50
60
Actual Demand Forecast
t
17ti
i1t Y7
1F
-
7/25/2019 ForecastMETHODS_TECHNIQUES
26/116
All Rights Reserved, Indian
Measures of aggregate error
Mean absolute error MAE
Mean absolute percentage
error MAPE
Mean squared error MSE
Root mean squared error
RMSE
n
ttE
nMAE
1
1
n
t t
t
Y
E
nMAPE
1
1
n
ttE
nMSE
1
21
n
ttE
nRMSE
1
21
Et = Ft - Yt
-
7/25/2019 ForecastMETHODS_TECHNIQUES
27/116
All Rights Reserved, Indian
DAD ForecastingMAPE and RMS
Mean absolute percentage
error MAPE
0.1068 or 10.68%
Root mean squared error
RMSE
5.8199
-
7/25/2019 ForecastMETHODS_TECHNIQUES
28/116
All Rights Reserved, Indian
-
7/25/2019 ForecastMETHODS_TECHNIQUES
29/116
All Rights Reserved, Indian
Exponential Smoothing
Form of Weighted Moving Average Weights decline exponentially.
Largest weight is given to the present observation, le
immediately preceding observation and so on.
Requires smoothing constant ()
Ranges from 0 to 1
-
7/25/2019 ForecastMETHODS_TECHNIQUES
30/116
All Rights Reserved, Indian
Exponential Smoothing
Next forecast = (present actual value) + (1-) prese
-
7/25/2019 ForecastMETHODS_TECHNIQUES
31/116
All Rights Reserved, Indian
Simple Exponential Smoothing Equati
Smoothing Equations
11
1 *)1(*
YF
FYF ttt
Ft+1 is the forecasted value at time t+1
-
7/25/2019 ForecastMETHODS_TECHNIQUES
32/116
All Rights Reserved, Indian
Simple Exponential Smoothing Equations
Smoothing Equations
....)1()1( 32
21 tttt YYYF
-
7/25/2019 ForecastMETHODS_TECHNIQUES
33/116
All Rights Reserved, Indian
Exponential Smoothing Forecast
0
10
20
30
40
50
60
18-09-2014 08-10-2014 28-10-2014 17-11-2014 07-12-2014 27-12-2014 16-01-2015 05-02-20
Demand Exponential Smoothing Forecast
ttt FYF *)8098.01(*8098.01
-
7/25/2019 ForecastMETHODS_TECHNIQUES
34/116
All Rights Reserved, Indian
DAD ForecastingMAPE and RMSE
Exponential Smoothing
Mean absolute percentage
error MAPE
0.0906 or 9.06%
Root mean squared error
RMSE
5.3806
= 0.8098
-
7/25/2019 ForecastMETHODS_TECHNIQUES
35/116
All Rights Reserved, Indian
Choice of
Forsmooth data, try a high value of a, forecast responsiv
current data.
Fornoisy data try low a forecast more stableless respon
-
7/25/2019 ForecastMETHODS_TECHNIQUES
36/116
All Rights Reserved, Indian
Optimal
11
1
2
)1(
/
ttt
n
t
tt
FYF
nFYMin
-
7/25/2019 ForecastMETHODS_TECHNIQUES
37/116
All Rights Reserved, Indian
-
7/25/2019 ForecastMETHODS_TECHNIQUES
38/116
All Rights Reserved, Indian
Double exponential Smoothing Holts m
Simple exponential smoothing may produce consistently biasthe presence of a trend.
Holt's method (double exponential smoothing) is appropriate has a trend but no seasonality.
Systematic component of demand = Level + Trend
-
7/25/2019 ForecastMETHODS_TECHNIQUES
39/116
-
7/25/2019 ForecastMETHODS_TECHNIQUES
40/116
All Rights Reserved, Indian
Holts Method
Holts Equations
Forecast Equation
11
11
)1()()(
)()1()(
tttt
tttt
TLLTii
TLYLi
ttt TLF
1
Equation for
intercept or
level
Equation for
Slope (Trend)
i i l l f d
-
7/25/2019 ForecastMETHODS_TECHNIQUES
41/116
All Rights Reserved, Indian
Initial values of Lt and Tt
L1is in general set to Y1.
T1can be set to any one of the following values (or use regression to ge
)1/()(
3/)()()()(
11
3423121
121
nYYT
YYYYYYTYYT
n
-
7/25/2019 ForecastMETHODS_TECHNIQUES
42/116
All Rights Reserved, Indian
0
10
20
30
40
50
60
0 20 40 60 80 100 120
Demand Forecast
Double exponential smoothing
= 0 8098; = 0 05
DAD Forecasting MAPE and RMSE
-
7/25/2019 ForecastMETHODS_TECHNIQUES
43/116
All Rights Reserved, Indian
DAD ForecastingMAPE and RMSE
Double Exponential Smoothing
Mean absolute percentage
error MAPE
0.0930 or 9.3%
Root mean squared error
RMSE
5.5052
= 0.8098; = 0.05
-
7/25/2019 ForecastMETHODS_TECHNIQUES
44/116
All Rights Reserved, Indian
Forecasting Accuracy
-
7/25/2019 ForecastMETHODS_TECHNIQUES
45/116
All Rights Reserved, Indian
Forecasting Accuracy
The forecast error is the difference between the forecast value and the a
the corresponding period.
tperiodfor timeForecastF
tperiodat timevalueActualY
tperiodaterrorForecastE
FYE
t
t
t
ttt
-
7/25/2019 ForecastMETHODS_TECHNIQUES
46/116
-
7/25/2019 ForecastMETHODS_TECHNIQUES
47/116
All Rights Reserved, Indian
Forecasting Power of a Model
Theils coefficient (U Statistic)
-
7/25/2019 ForecastMETHODS_TECHNIQUES
48/116
All Rights Reserved, Indian
Theils coefficient (U Statistic)
n
ttt
n
t tt
YY
FYU
1
21
1
2
11
)(
)(
The value U is the relative forecasting power of the method against nave tec
the technique is better than nave forecast If U > 1, the technique is no bettforecast.
-
7/25/2019 ForecastMETHODS_TECHNIQUES
49/116
All Rights Reserved, Indian
Theils coefficient for DAD Hos
Method U
Moving Average with 7 periods 1.221
Exponential Smoothing 1.704
Double exponential Smoothing 1.0310
Th il C ffi i t
-
7/25/2019 ForecastMETHODS_TECHNIQUES
50/116
All Rights Reserved, Indian
Theils Coefficient
1-n
1t
2
1
21-n
1t
11t
1
2
1
2
1
2 F
U2,1
t
tt
t
t
n
t
t
n
t
t
n
t
tt
Y
YY
Y
Y
FY
FYU
U1 is bounded between 0 and 1, with values closure to zero indicating
IfU 2 = 1, there is no difference between nave forecast and the foreca
IfU 2 < 1, the technique is better than nave forecast
IfU 2 > 1, the technique is no better than the nave forecast.
-
7/25/2019 ForecastMETHODS_TECHNIQUES
51/116
All Rights Reserved, Indian
S l Eff t
-
7/25/2019 ForecastMETHODS_TECHNIQUES
52/116
All Rights Reserved, Indian
Seasonal Effect
Seasonal effect is defined as the repetitive and predictable pbehaviour in a time-series around the trend line.
In seasonal effect the time period must be less than one year,
weeks, months or quarters.
Seasonal effect
-
7/25/2019 ForecastMETHODS_TECHNIQUES
53/116
All Rights Reserved, Indian
Seasonal effect
Identification of seasonal effect provides better understandinseries data.
Seasonal effect can be eliminated from the time-series. This prdeseasonalization or seasonal adjusting.
-
7/25/2019 ForecastMETHODS_TECHNIQUES
54/116
All Rights Reserved, Indian
Seasonal Adjusting
tttt
tt
tt
TSTY
modeladditiveinadjustmentSeasonal
100TT
STeffectSeasonal
modetivemultiplicainadjustmentSeasonal
tt
t
SS
Y
Seasonal Index
-
7/25/2019 ForecastMETHODS_TECHNIQUES
55/116
All Rights Reserved, Indian
Seasonal Index
Method of simple averages
Ratio-to-moving average method
Method of simple averages
-
7/25/2019 ForecastMETHODS_TECHNIQUES
56/116
All Rights Reserved, Indian
Method of simple averages
Average the unadjusted data by period ( for example daily or month
Calculate the average of daily (or monthly) averages.
Seasonal index for day i (or month i) is the ratio of daily avera
month i) to the average of daily (or monthly) averages times 100.
Example: DAD Exa
mple - Demand for Continental Bre
-
7/25/2019 ForecastMETHODS_TECHNIQUES
57/116
All Rights Reserved, Indian
Example: DAD Example - Demand for Continental Bre
5 October1 November 2014 Data
DAY
Week
(5-11 October)
Week 2
(12-18 October)
Week 3
(19-25 OCT)
Sunday 41.00 40.00 40.00
Monday 30.00 44.00 43.00
Tuesday 40.00 49.00 41.00
Wednesday 40.00 50.00 46.00
Thursday 40.00 45.00 41.00
Friday 40.00 40.00 40.00
Saturday 40.00 42.00 32.00
Seasonality Index
-
7/25/2019 ForecastMETHODS_TECHNIQUES
58/116
All Rights Reserved, Indian
Seasonality Index
DAY
Week1
(5-11 October)
Week 2
(12-18 October)
Week 3
(19-25 OCT)
Week 4
(26 Oct-1 Nov)
Daily
Average
Sunday 41.00 40.00 40.00 41.00 40.50
Monday 30.00 44.00 43.00 41.00 39.50
Tuesday 40.00 49.00 41.00 40.00 42.50
Wednesday 40.00 50.00 46.00 43.00 44.75
Thursday 40.00 45.00 41.00 46.00 43.00
Friday 40.00 40.00 40.00 45.00 41.25
Saturday 40.00 42.00 32.00 45.00 39.75
41.61
Average of daily
averages
Deseasonalized Data
-
7/25/2019 ForecastMETHODS_TECHNIQUES
59/116
All Rights Reserved, Indian
Deseasonalized Data
Date Demand Seasonal Index De-seasonalized Data
10/5/2014 41 97.34% 42.12
10/6/2014 30 94.94% 31.60
10/7/2014 40 102.15% 39.16
10/8/2014 40 107.55% 37.19
10/9/2014 40 103.34% 38.70
10/10/2014 40 99.14% 40.3510/11/2014 40 95.54% 41.87
Trend
-
7/25/2019 ForecastMETHODS_TECHNIQUES
60/116
All Rights Reserved, Indian
Trend
Trend is calculated using regression on deseasonalized d
Deseasonalized data is obtained by dividing the actual
seasonality index.
-
7/25/2019 ForecastMETHODS_TECHNIQUES
61/116
All Rights Reserved, Indian
SUMMARY OUTPUT
Regression Statistics
Multiple R 0.268288
R Square 0.071978Adjusted R Square 0.036285
Standard Error 3.715395
Observations 28
ANOVA
df SS MS F Significa
Regression 1 27.83729 27.83729 2.016587 0.1674
Residual 26 358.9082 13.80416
Total 27 386.7455
Coefficients Standard Error t Stat P-value Lower 9
Intercept 39.81731 1.442768 27.59786 8.7E-21 36.851
Day 0.123437 0.086923 1.420066 0.167469 -0.055
Forecast
-
7/25/2019 ForecastMETHODS_TECHNIQUES
62/116
All Rights Reserved, Indian
Ft = Tt x St
F29 = T29 x S29
F29 = (39.81 + 0.1234 x 29) x 0.9733 = 42.2372
Y29 = 46
-
7/25/2019 ForecastMETHODS_TECHNIQUES
63/116
-
7/25/2019 ForecastMETHODS_TECHNIQUES
64/116
All Rights Reserved, Indian
-
7/25/2019 ForecastMETHODS_TECHNIQUES
65/116
A t l ti F ti (ACF)
-
7/25/2019 ForecastMETHODS_TECHNIQUES
66/116
All Rights Reserved Indian
Auto-correlation Function (ACF)
A k-period plot of autocorrelations is called autofunction (ACF) or a correlogram.
Auto-Correlation
-
7/25/2019 ForecastMETHODS_TECHNIQUES
67/116
All Rights Reserved Indian
Auto-correlation of lag k is auto-correlation between
To test whether the autocorrelation at lag k is significa
from 0, the following hypothesis test is used:
H0: k = 0
HA: k 0
For any k, reject H0 if | k| > 1.96/n. Where n is thobservations.
ACF for Demand for Continental
B kf t
-
7/25/2019 ForecastMETHODS_TECHNIQUES
68/116
All Rights Reserved Indian
Breakfast
Partial Auto-correlation
-
7/25/2019 ForecastMETHODS_TECHNIQUES
69/116
All Rights Reserved, Indian
Partial auto correlation of lag k is an auto correlation betweewith linear dependence between the intermedia values (Yt-k+1, Yremoved.
Partial Auto correlation Function
-
7/25/2019 ForecastMETHODS_TECHNIQUES
70/116
All Rights Reserved, Indian
Partial Auto-correlation Function
A k-period plot of partial autocorrelations is caautocorrelation function (PACF).
Partial Auto-Correlation
-
7/25/2019 ForecastMETHODS_TECHNIQUES
71/116
All Rights Reserved, Indian
Partial auto-correlation of lag k is auto-correlation between Yt aremoval of linear dependence of Yt+1 to Yt+k-1.
To test whether the partial autocorrelation at lag k is significantlythe following hypothesis test is used:
H0:pk = 0
HA:pk 0
For any k, reject H0 if |pk| > 1.96/n. Where n is the number of
PACF Demand for Continental Breakfast at
-
7/25/2019 ForecastMETHODS_TECHNIQUES
72/116
All Rights Reserved, Indian
PACF Demand for Continental Breakfast at
DAD hospital
Stationarity
-
7/25/2019 ForecastMETHODS_TECHNIQUES
73/116
All Rights Reserved, Indian
A time series is stationary, if:
Mean () is constant over time.
Variance () is constant over time.
The covariance between two time periods (Yt) and (only on the lag k not on the time t.
We assume that the time series is stationary before
applying forecasting models
ACF Plot of non-stationary and stationary
-
7/25/2019 ForecastMETHODS_TECHNIQUES
74/116
All Rights Reserved, Indian
process
Non-stationary Stationary
White Noise
-
7/25/2019 ForecastMETHODS_TECHNIQUES
75/116
All Rights Reserved, Indian
White noise is a data uncorrelated across time that fodistribution with mean 0 and constant standard deviat
In forecasting we assume that the residuals are white N
-
7/25/2019 ForecastMETHODS_TECHNIQUES
76/116
All Rights Reserved, Indian
Residual White Noise
-
7/25/2019 ForecastMETHODS_TECHNIQUES
77/116
All Rights Reserved, Indian
Auto-Regression
-
7/25/2019 ForecastMETHODS_TECHNIQUES
78/116
All Rights Reserved, Indian
Auto-regression is a regression of Y on itself observed
time points.
That is, we use Yt as the response variable and Yt-1, Yt-2explanatory variables.
AR(1) Parameter Estimation
-
7/25/2019 ForecastMETHODS_TECHNIQUES
79/116
All Rights Reserved, Indian
n
t
t
n
t
tt
n
t
tt
n
t
t
ttt
Y
YY
YY
YY
2
21
2
1
2
21
2
2
1
OLS
Estimate
AR(1) Process
-
7/25/2019 ForecastMETHODS_TECHNIQUES
80/116
All Rights Reserved, Indian
Yt = Yt-1 + t
Yt = (Yt-2 + t-1) + t
Yt = t X0 +
t-1 1 + t-2 2++
1 t-1 + t
1
0
0
t
i
itit
t XY
Auto-regressive process (AR(p))
-
7/25/2019 ForecastMETHODS_TECHNIQUES
81/116
All Rights Reserved, Indian
Assume {Yt} is purely random with mean zero and consdeviation (White Noise).
Then the autoregressive process of orderp or AR(p) proc
tptpttt YYYY ...22110
AR(p) process models each future observationas a function p previous observations.
-
7/25/2019 ForecastMETHODS_TECHNIQUES
82/116
All Rights Reserved, Indian
Model Identification in AR(p)Process
Pure AR Model Identification
-
7/25/2019 ForecastMETHODS_TECHNIQUES
83/116
All Rights Reserved, Indian
Model ACF PACF
AR(1) Exponential Decay: Positive side if 1
> 0 and alternating in sign starting onnegative side if 1 < 0.
Spike at lag 1, then cuts off t
Spike positive if 1 > 0 and nside if 1 < 0.
AR(p) Exponential decay: pattern depends onsigns of 1, 2, etc
Spikes at lags 1 to p, then cutzero.
tptpttt YYYY ...22110
Partial Auto-Correlation
-
7/25/2019 ForecastMETHODS_TECHNIQUES
84/116
All Rights Reserved, Indian
Partial auto-correlation of lag k is auto-correlation between Yt anremoval of linear dependence of Yt+1 to Yt+k-1.
To test whether the autocorrelation at lag k is significantly differfollowing hypothesis test is used:
H0: k = 0
HA: k 0
For any k, reject H0 if | k| > 1.96/n. Where n is the number of
PACF FunctionDAD ContinentaBreakfast
-
7/25/2019 ForecastMETHODS_TECHNIQUES
85/116
All Rights Reserved, Indian
First t
PAC a
differe
from
zero.
ACF Function - DAD Continental
Breakfast
-
7/25/2019 ForecastMETHODS_TECHNIQUES
86/116
All Rights Reserved, Indian
Actual Vs Fit- Continental Breakfas
Data
-
7/25/2019 ForecastMETHODS_TECHNIQUES
87/116
All Rights Reserved, Indian
Data
AR(2) Model
-
7/25/2019 ForecastMETHODS_TECHNIQUES
88/116
All Rights Reserved, Indian
MAPE = 8.8%
Residual White Noise
-
7/25/2019 ForecastMETHODS_TECHNIQUES
89/116
All Rights Reserved, Indian
-
7/25/2019 ForecastMETHODS_TECHNIQUES
90/116
All Rights Reserved, Indian
(Yt 41.252) = 0.663 (Yt-1 41.252) 0.204 (Yt-2 41.252)
Yt = (Xt - )
-
7/25/2019 ForecastMETHODS_TECHNIQUES
91/116
All Rights Reserved, Indian
Moving Average Process
-
7/25/2019 ForecastMETHODS_TECHNIQUES
92/116
All Rights Reserved, Indian
A moving average process is a time series regression modelthe value at time t, Yt, is a linear function of past errors.
First order moving average, MA(1), is given by:
Yt = 0 + 1t-1 + t
Moving Average Process MA(q)
-
7/25/2019 ForecastMETHODS_TECHNIQUES
93/116
All Rights Reserved, Indian
{Yt} is a moving average process of order q (written MA(q)) if
constants 0, 1, . . . q
We have.,
tqtqtttY ...2210
MA(q) models each future observation as a function ofq previous errors
Pure MA Model Identification
-
7/25/2019 ForecastMETHODS_TECHNIQUES
94/116
All Rights Reserved, Indian
Model ACF PACF
MA(1) Spike at lag 1 then cuts of to zero. Spikepositive if 1 > 0 and negative side if
1 < 0.
Exponential decay. On neg1> 0 on positive side i
MA(q) Spikes at lags 1 to q, then cuts off to zero. Exponential decay or sine w
pattern depends on signs o
ACF Function - DAD ContinentalBreakfast
-
7/25/2019 ForecastMETHODS_TECHNIQUES
95/116
All Rights Reserved, Indian
SPSS Output
-
7/25/2019 ForecastMETHODS_TECHNIQUES
96/116
All Rights Reserved, Indian
Actual Vs Fitted Value
-
7/25/2019 ForecastMETHODS_TECHNIQUES
97/116
All Rights Reserved, Indian
Residual Plot
-
7/25/2019 ForecastMETHODS_TECHNIQUES
98/116
All Rights Reserved, Indian
-
7/25/2019 ForecastMETHODS_TECHNIQUES
99/116
All Rights Reserved, Indian
-
7/25/2019 ForecastMETHODS_TECHNIQUES
100/116
ARMA(
p,q) Model Identification
-
7/25/2019 ForecastMETHODS_TECHNIQUES
101/116
All Rights Reserved, Indian
ARMA(p,q) models are not easy to identify. We usually start wAR and MA process. The following thump rule may be used.
The final ARMA model may be selected based on parameters sRMSE, MAPE, AIC and BIC.
Process ACF PACF
ARMA(p,q) Tails off after q lags Tails off to zero after p
lags
-
7/25/2019 ForecastMETHODS_TECHNIQUES
102/116
All Rights Reserved, Indian
ACF PACF
ARMA(2,1) Model Output
-
7/25/2019 ForecastMETHODS_TECHNIQUES
103/116
All Rights Reserved, Indian
AR(2) Model
-
7/25/2019 ForecastMETHODS_TECHNIQUES
104/116
All Rights Reserved, Indian
MAPE = 8.8%
Residual White Noise
-
7/25/2019 ForecastMETHODS_TECHNIQUES
105/116
All Rights Reserved, Indian
Forecasting Model Evaluation
Akaikes information criteria
-
7/25/2019 ForecastMETHODS_TECHNIQUES
106/116
All Rights Reserved, Indian
Akaike s information criteria
AIC = -2LL + 2m
Where m is the number of variables estimated in the model
Bayesian Information Criteria
BIC = -2LL + m ln(n)
Where m is the number of variables estimated in the model and n is the numbeobservations
Final Model Selection
-
7/25/2019 ForecastMETHODS_TECHNIQUES
107/116
All Rights Reserved, Indian
Model AR(2) MA(1) ARMA(2,1)
BIC 3.295 3.301 3.343
AR(2) has the
least BIC
-
7/25/2019 ForecastMETHODS_TECHNIQUES
108/116
All Rights Reserved, Indian
ARIMA
ARIMA has the following three components:
-
7/25/2019 ForecastMETHODS_TECHNIQUES
109/116
All Rights Reserved, Indian
ARIMA has the following three components:
Auto-regressive component: Function of past values of the time
Integration Component: Differencing the time series to make it a st
Moving Average Component: Function of past error values.
Integration (d)
-
7/25/2019 ForecastMETHODS_TECHNIQUES
110/116
All Rights Reserved, Indian
Used when the process is non-stationary.
Instead of observed values, differences between observed values are
When d=0, the observations are modelled directly. If d = 1,between consecutive observations are modelled. If d = 2, the dif
differences are modelled.
ARIMA (p, d, q)
-
7/25/2019 ForecastMETHODS_TECHNIQUES
111/116
All Rights Reserved, Indian
The q and p values are identified using auto-correlation
and Partial auto-correlation function (PACF) respectivelidentifies the level of differencing.
Usually p+q
-
7/25/2019 ForecastMETHODS_TECHNIQUES
112/116
All Rights Reserved, Indian
Differencing is a process of making a non-stationary
stationary process.
In differencing, we create a new process Xt, where Xt = Yt
ARIMA(p,1,q) Process
-
7/25/2019 ForecastMETHODS_TECHNIQUES
113/116
All Rights Reserved, Indian
tqtqtt
ptpttt
XXXX
...
...
22110
22110
Where Xt = YtYt-1
Ljung-Box Test
Ljung-Box is a test on the autocorrelations of residuals. The test sta
-
7/25/2019 ForecastMETHODS_TECHNIQUES
114/116
All Rights Reserved, Indian
m
k
k
m kn
rnnQ
1
2
)2(
n = number of observations in the time series.
k = particular time lag checkedm = the number of time lags to be tested
rk = sample autocorrelation function of the kth residual term.
Q statistic is approximate chi-square distribution with mpq degrees of freedom are correctly specified.
H0: The model does not exhibit lack of fit
HA: The model exhibits lack of fit
AR(2)
Ljung-Box Test
-
7/25/2019 ForecastMETHODS_TECHNIQUES
115/116
All Rights Reserved, Indian
P-value is 0.740, thus the model doesnt show
lack of fit
Box-Jenkins Methodology
Identification: Identify the ARIMA model using ACF & PAC
-
7/25/2019 ForecastMETHODS_TECHNIQUES
116/116
All Rights Reserved, Indian
y gwould give the values of p, q and d.
Estimation: Estimate the model parameters (using maximum
Diagnostics: Check the residual for any issue such as not pr
Noise.