Week 9 - Forecasting 1

Question 1

Why is forecasting important? (2)

Answer 1

• Are an important factor in enabling organisations to deliver goods and services to the customer when required and thus achieve a quality service
• Forecasting enables us to plan for production, transportation and outsourcing and ensuring appropriate levels of capacity and inventory

Question 2

What is forecast accuracy dependant on? (2)

Answer 2

• Collecting up-to-date data on relevant information such as prices and sales volume and choose an appropriate forecasting technique
• The time time horizon over which it is serviced - forecasts for short term horizons tend to be more accurate than longer term forecasts

Question 3

Which two ways can forecast be developed?

Answer 3

• Quantitative - Time series, causal
• Qualitative - Survey, Delphi, expert

Question 4

What is qualitative forecasting?

Answer 4

They take a subjective approach and are based on estimated and opinions

Question 5

3 techniques of qualitative techniques

Answer 5

• Market surveys
• Delphi method
• Expert judgement

Question 6

What is a market survey? (3)

Answer 6

• It collects data from a sample of customers, analyses the responses and makes inferences about the population from the sample
• Useful before launch of a new product when there is little info on potential customer demand
• It can be expensive and time-consuming

Question 7

What is a Delphi study? (4)

Answer 7

• A questionnaire is completed by a panel of experts which is analysed and summarised
• Each expert can compare their forecast with the summarised reply of the others
• This repeated until a consensus has emerged within the group on the best decision
• Good accuracy but costly and high effort

Question 8

What is expert judgement? (4)

Answer 8

• Can be either individual or group judgement
• Is unreliable and risky
• A group judgement relies on a consensus being found which is more reliable than an individual
• Cheap method

Question 9

What are the four components of a time series?

Answer 9

• Trend
• Cyclical
• Seasonal
• Irregular

Question 10

What is a trend? (4 components of a time series)

Answer 10

A gradual movement to relatively higher or lower values over time (not random fluctuations)

Question 11

When is something cyclical? (4 components of a time series)

Answer 11

Any recurring sequence of points above and below the trend line lasting for more than one year

Question 12

What is meant by seasonal? (4 components of a time series)

Answer 12

Any regularly repeating pattern that is less than one year in duration

Question 13

What is meant by irregular? (4 components of a time series)

Answer 13

Deviations from the time series values from those expected by the trend, cyclical and seasonal component (cannot predict its impact on the time series)

Question 14

When is time series decomposition appropriate?

Answer 14

If seasonal variation is evident in the demand pattern and the effect of seasonality is to be included in the forecast

Question 15

What is time series decomposition?

Answer 15

It consists of smoothing out past values to eliminate randomness so that the pattern and the effect of seasonality is to be included in the forecast

Question 16

Two common forms of the time series that can be used for decomposition

Answer 16

Additive form and the multiplicative form

Question 17

When is additive model (form) appropriate

Answer 17

If the magnitude of the seasonal fluctuations doesn’t vary with the level of series

Question 18

When is multiplicative decomposition more useful?

Answer 18

In economic series as more seasonal economic series do have seasonal variation, which increases with the level of the series

Question 19

Additive form equation

Answer 19

Question 20

Multiplicative form equation

Answer 20

Question 21

Steps to decomposing a series (4)

Answer 21

• Use a 4 period moving average
• Assuming a multiplicative model, moving average requires an odd number of observations to ensure the average is centred at the middle of the data values being averaged
• Find centred moving average by doing averages of 1-4 and 2-5 then find the average of that
• Then divide the observation by the centred moving average to identify the seasonal-irregular value
• Then find the seasonal index for each quarter by getting the average

Question 22

How to calculate the adjusted seasonal index?

Answer 22

Multiply each seasonal index by the number of seasons divided by the unadjusted seasonal indexes

Question 23

How to calculate deseasonalised figures

Answer 23

In a multiplicative model we divide each time series observation by the corresponding seasonal index