Forecasting

Natasya Eveline
15

People i Gede Dipta Mahesa

behind 14

Ni Ketut Mas Ayu Anditi

34
What Is
Forecasting?
Forecasting is the art and science of
predicting future events. Forecasting
may involve taking historical data such
as past sales and projecting them into
the future with a mathematical model.
It may be a subjective or an intuitive
prediction. It may be based on
demand-driven data, such as customer
plans to purchase, and projecting them
into the future. Or the forecast may
involve a combination of these, that is,
a mathematical model adjusted by a
manager’s good judgment.
Time Types of
Horizons Forecast
Short- range forecast
Economic forecasts address the business
less than 3 month
cycle by predicting inflation rates, money
supplies, housing starts, and other
planning indicators.
Medium-range forecast
3 month - 3 years
Technological forecasts are concerned with
rates of technological progress, which can
Long-range forecast result in the birth of exciting new products,
3 years or more lifespan requiring new plants and equipment.

Demand forecasts are projections of demand for

a company’s products or services. Forecasts
drive decisions, so managers need immediate
and accurate information about real demand.
7stepsin
forecasting
system Determine the Select the Determine the
use of the items to be time horizon of
forecast. forecasted. the forecast.

Select the Gather the

data needed to Make the Validate and
forecasting forecast. implement the
model. make the
forecast. results.
7stepsin
forecasting
system Determine the Select the Determine the
use of the items to be time horizon of
forecast. forecasted. the forecast.

Select the Gather the

data needed to Make the Validate and
forecasting forecast. implement the
model. make the
forecast. results.
 ForecastingApproaches
Quantitative forecasts Quantitative forecasts
Forecasts that employ mathematical modeling to Forecasts that employ mathematical
forecast demand. modeling to forecast demand.

Methods: Methods:
1. Jury of executive opinion A forecasting 1. Naive approach
technique that uses the opinion of a small 2. Moving averages Time-series
group of high-level managers to form a group 3. Exponential smoothing models
estimate of demand 4. Trend projection
Associative
2. Delphi method A forecasting technique using 5. Linear regression model
a group process that allows experts to make
forecasts. Time-Series Models predict on the
3. Sales force composite A forecasting technique assumption that the future is a function of
based on salespersons’ estimates of expected the past.
sales. Associative Models such as linear
4. Market survey A forecasting method that regression, incorporate the variables or
solicits input from customers or potential factors that might influence the quantity
customers regarding future purchasing plans. being forecast.
Time-Series Forecasting
Time series forecasting is a quantitative method used to predict future
values of a variable based on its historical patterns. It involves
analyzing the past data points of a variable and identifying any trends,
seasonal patterns, or cyclical components present in the data. These
components are then used to forecast future values of the variable.
 Decomposition of Time Series
Analyzing time series means breaking down past data into components
and then projecting them forward. A time series has four components:

1. Trend is the gradual upward or downward movement of the data over

time. Changes in income, population, age distribution, or cultural views may
account for movement in trend.

2. Seasonality is a data pattern that repeats itself after a period of days,

weeks, months, or quarters. There are six common seasonality patterns:
 Decomposition of Time Series
3. Cycles are patterns in the data that occur every several years. They are
usually tied into the business cycle and are of major importance in short-
term business analysis and planning. Predicting business cycles is difficult
because they may be affected by political events or by international turmoil.

4. Random variations are “blips” in the data caused by chance and unusual
situations. They follow no discernible pattern, so they cannot be predicted.
 Naive Forecasting Method

The simplest forecasting

method assumes that the
demand in the next period will
be equal to the demand in the
most recent period. . In other
words, if sales of a product—
say, Nokia cell phones—were 68
units in January, we can
forecast that February’s sales
will also be 68 phones.

Forecast for the next period = Actual demand in the most recent period
 Moving Averages
Moving averages are used to
smooth out fluctuations in the data
by averaging a specified number of
past observations. Moving averages
are useful if we can assume that
market demands will stay fairly
steady over time .

Simple Moving Average (SMA) =

(Sum of the last n observations) / n

Where n is the number of periods in

the moving average.
 Weighted Moving
Averages
Weighted moving averages assign
more weight to recent observations
and less weight to older
observations.

Weighted Moving Average =

 Exponential Smoothing
Exponential smoothing is a weighted moving average method that assigns exponentially
decreasing weights to older observations. Eksponential smoothing can written
mathematically as

Exponential smoothing =
 Measuring Forecast Error

The overall accuracy of any forecasting model—moving average, exponential smoothing,

or other—can be determined by comparing the forecasted values with the actual or
observed values. If F t denotes the forecast in period t , and A t denotes the actual
demand in period t , the forecast error (or deviation) is defined as: Forecast error =
Actual demand - Forecast value = At - Ft. These measures can be used to compare
different forecasting models, as well as to monitor forecasts to ensure they are
performing well. Three of the most popular measures are mean absolute deviation
(MAD), mean squared error (MSE), and mean absolute percent error (MAPE).
Measuring Forecast Error
 Exponential Smoothing with Trend Adjustment
Exponential smoothing fails to respond to trends in the data. To address this, the trend-adjusted
exponential smoothing technique is used. The idea is to compute an exponentially smoothed
average of the data and then adjust for positive or negative trend.
The formula for the forecast including trend (FIT) is:
FITt = Ft + Tt
With trend-adjusted exponential smoothing, estimates for both the average and the trend are
smoothed. This procedure requires two smoothing constants: a for the average and b for the trend.
We then compute the average and trend each period:
Ft = a(Actual demand last period) + (1 - a)(Forecast last period + Trend estimate last period)
or: Ft = a(At - 1) + (1 - a)(Ft - 1 + Tt - 1)

Tt = b(Forecast this period - Forecast last period) + (1 - b)(Trend estimate last period) or: Tt =
b(Ft - Ft - 1) + (1 - b)Tt - 1

where
Ft = exponentially smoothed forecast average of the data series in period t
Tt = exponentially smoothed trend in period t
At = actual demand in period t a = smoothing constant for the average (0 a 1)
b = smoothing constant for the trend (0 b 1)
 Exponential Smoothing with Trend Adjustment
So the three steps to compute a trend-adjusted forecast are:
STEP 1: Compute F t , the exponentially smoothed forecast average for period t , using Equation.
STEP 2: Compute the smoothed trend, T t , using Equation .
STEP 3: Calculate the forecast including trend, FIT t , by the formula FIT t = F t + T t.
Exponential Smoothing with Trend Adjustment
 Trend Projection
Trend projection is a time-series forecasting technique that fits a trend line to a series of
historical data points and then projects the line into the future to make forecasts, typically for
the medium to long-range. While several mathematical trend equations can be used (e.g.
exponential, quadratic), this section focuses on linear (straight-line) trends using the least-
squares method.
yn = a + bx
Where:
yn = computed value of the variable to be predicted (dependent variable)
a = y-intercept
b = slope of the regression line (rate of change in y for given change in x)
x = independent variable, which is time in this case

The formulas to calculate the slope b and y-intercept a are:

b = Σ(xy) - nxy / Σ(x^2) - nx^2
a = ybar - b(bar)

Where: Σ = summation x, y = known values of independent and dependent variables xbar,

ybar = means of x and y values n = number of data points
 Seasonal variations
Seasonal variations are regular movements in a time series related to recurring events such as
weather or holidays. Examples include higher demand for heating fuels in winter months, golf
clubs or sunscreen in summer months. Analyzing monthly or quarterly data makes it easier to
spot seasonal patterns. The steps to forecast using a multiplicative seasonal model are:
1. Calculate the average historical demand for each season/month
2. Calculate the average demand across all seasons/months
3. Compute a seasonal index by dividing (1) by (2)
4. Estimate next year's total annual demand
5. Divide (4) by the number of seasons, then multiply by the seasonal index to get the seasonal
forecast

Cycles are similar to seasonal variations, but they occur over several years rather than weeks,
months or quarters. Forecasting cyclical variations in a time series is difficult. Cycles involve a
wide variety of factors that cause the economy to fluctuate between periods of recession and
expansion over several years. These factors include overexpansion across the nation or an
industry during times of economic euphoria, followed by contraction during periods of economic
concern.
Associative Forecasting Methods: Regression and
Correlation Analysis

consider several variables that are related to the quantity being

predicted. Once these related variables have been found, a
statistical model is built and used to forecast the item of interest.
This approach is more powerful than the time-series methods that
use only the historical values for the forecast variable.
Using Regression Analysis for Forecasting
Nodel Construction Company renovates old homes in West Bloomfield, Michigan. Over time, the com-
pany has found that its dollar volume of renovation work is dependent on the West Bloomfield area
payroll. Management wants to establish a mathematical relationship to help predict sales.
 If the local chamber of commerce predicts
that the West Bloomfield area payroll will be
$6 billion next
year, we can estimate sales for Nodel with the
regression equation:

Sales (in + millions)

= 1.75 + .25(6)
= 1.75 + 1.50
= 3.25
 Standard Error of the Estimate

To measure the accuracy of the regression estimates, we must compute the standard error of the estimate,
Sy, x. This computation is called the standard deviation of the regression: It mea- sures the error from the
dependent variable, y, to the regression line, rather than to the mean.
 Correlation Coefficients for Regression Lines

The regression equation is one way of expressing the nature of the relationship between two variables.
The regression equation shows how one variable relates to the value and changes in another variable.
Another way to evaluate the relationship between two variables is to compute the coefficient of correlation.
This measure expresses the degree or strength of the linear relationship
 Usually identified as r, the coefficient of correlation can be any number between + 1 and - 1.
To compute r, we use much of the same data needed earlier to calculate a and b for the regression line. The
rather lengthy equation for r is:
 Multiple-Regression Analysis
An associative forecasting method with more than one independent variable.
 Nodel Construction wants to see the impact of a second independent
variable, interest rates, on its sales.

E We also find that the new coefficient of correlation is .96, implying the
X inclusion of the variable x2, inter- est rates, adds even more strength to
the linear relationship.
A
M We can now estimate Nodel’s sales if we substitute values for next year’s payroll and
P interest rate. If West Bloomfield’s payroll will be $6 billion and the interest rate will be .12
(12%), sales will be forecast as:
L
E
 One way to monitor forecasts to
ensure that they are performing well is
to use a tracking signal. A tracking
signal is a measurement of how well a
forecast is predicting actual values. As
forecasts are updated every week,
month, or quarter, the newly available
Monitoring and demand data are com- pared to the

Controlling Forecasts forecast values.

 TO FIND MAD :

Positive tracking signals indicate that demand is greater than forecast. Negative signals
mean that demand is less than forecast.

In other words, small deviations are okay, but positive and negative errors should
balance one another so that the tracking signal centers closely around zero.

A consistent tendency for forecasts to be greater or less than the actual values (that is,
for a high absolute cumulative error) is called a bias error.
Once tracking signals are calculated, they are compared with predetermined control limits.
When a tracking signal exceeds an upper or lower limit, there is a problem with the
forecasting method, and management may want to reevaluate the way it forecasts
demand.

How do firms decide what the upper and lower tracking limits should be? There is no single
answer, but they try to find reasonable values—in other words, limits not so low as to be
triggered with every small forecast error and not so high as to allow bad forecasts to be
regularly overlooked.
 COMPUTING THE TRACKING SIGNAL AT CARLSON’S BAKERY

Carlson’s Bakery wants to evaluate performance of its croissant forecast.

with Develop a tracking signal for the forecast, and see if it stays within acceptable limits,

Because the tracking signal drifted from -2 MAD to +2.5 MAD (between 1.6 and 2.0
standard deviations), we can conclude that it is within acceptable limits.
 Adaptive Smoothing Focus Forecasting

An approach to exponential Forecasting that tries a variety of

smoothing forecasting in which the computer models and selects the
smoothing constant is automati- best one for a particular
cally changed to keep errors to a application.
minimum.
ForecastingintheService
Sector
Presents Unusual challenges
Special needs for short-term records
Needs differ greatly as function of industry and
product
Holidays and other calendar events
Unusual events

Example : Fastfood-restaurant, Specialty reatil

shop, Staffing for hospitals, Bangking, etc.
thank
you