W7

Question 1

forecasting across the organisation

Answer 1

• predicting possible future outcomes using qualitative and quantitative
• more accurate for grouped data - having a group of student who do different subjects is stronger than one student who does just one subject
• forecasts are more accurate for shorter than longer periods of time

Question 2

forecasting steps

Answer 2

1. deciding what needs to be forecast
2. evaluate and analyse appropriate data - identify needed data and whether it is available
3. select and test the forecasting model - cost, ease and use of accuracy
4. generates the forecast
5. monitor forecast accuracy over time - is there any randomisation?

Question 3

types of forecasts

Answer 3

qualitative - subjective
- forecasts generated subjectively by the forecaster
- educated guesses

quantitative methods - objective
- forecasts generated through mathematical modelling

• Having access to the right data takes a lot of time and can be costly as well (disad for quantitative)
• Can be biased when looking at qualitative point of view (disad of qual)

Question 4

qualitative method

Answer 4

• difference between executive opinion and Delphi method
• Delphi - group does not know they are being used to forecast data
• executive - are aware they are in a group used for data

Question 5

quantitative methods

Answer 5

• 2 models

1. time series model - your past predicts your future
2. causal model - explores cause and effect

Question 6

time series model

Answer 6

• two key things historic patterns and random variation
• need to find historic pattern - specific pattern in the past which will help predict the future

1. level - data fluctuates around a constant mean over time
2. trend - upward trend or downward trend over time
3. seasonality - any pattern that regularly repeats itself and is of a constant length
4. cycle - patterns created by economic fluctuations based on economical situations

(look at diagram)

• random variations - cannot predict or account for these variations

Question 7

time series models:

Answer 7

• naive - forecasted value for the next period is the same as the current time period (look at equation)
• simple mean - the average of all available data (look at equation)
• actual values / number of observations
• moving average - the average value over a set time period / each new forecast drops the oldest data point and adds a new observation (look at equation)
• weighted moving average - weight should add up to 1 / allows emphasising one period over others (look at equation)

Question 8

exponential smoothing

Answer 8

• need three pieces of data to start
  1. last period forecast (Ft)
  2. last periods actual value (At)
  3. select value of smoothing coefficient - between 0 and 1 (alpha)
• If no last period forecast is available, average the last few periods or use naïve method
• Higher alpha values (e.g 0.7 or 0.8) may place too much weight on last periods random variation
  ○ Try not to use a big value
  ○ If you use bigger smoothing you place a lot of emphasis on the actual value rather than the forecasted value
  Smaller is better as want to place emphasis on forecasted value not actual value
• exponential equation = alpha value x actual value in current period + (1-alpha) x forecasted value

Question 9

forecasting seasonality

Answer 9

• Calc the avg. demand per season
  ○ E.g average quarterly demand
• Calculate seasonal index
  ○ Divide the actual demand of each season by the average demand per season for that year
• Avg. of the index
  ○ E.g - take the average of all spring indexes, then all summer indexes etc..
• forecast demand for the next year and divide by the number of seasons
  ○ Use regular forecasting method and divide by four for average quarterly demand
• Multiple next years average seasonal demand by each average seasonal index
  Result is a forecast of demand for each season of next year

Question 10

selecting the right forecasting model

Answer 10

• The amount and type of available data
  ○ Some methods require more data than others
• Degree of accuracy required
  ○ Increasing accuracy means more data
• Length of forecast horizon
  ○ Different models for 3 month vs 10 years
• Presence of data patterns
  Lagging will occur when a forecasting model meant for a level pattern is applied with a trend