A Pull Planning Framework - Forecasting
9 important questions on A Pull Planning Framework - Forecasting
What is the difference between qualitative and quantitative forecasting?
Qualitative forecasting is forecasting by using the expertise of people.
Quantitative forecasting is forecasting by using numerical measures in some kind of mathematical model.
Note: the book says: quantitative forecasting is forecasting by using numerical measures text-decorationof the past in some kind of mathematical model. This is not always true for causal models.
What's the difference between causal forecasting models and time series forecasting models?
Causal forecasting models predict a future parameter as a function of other parameters.
Time series forecasting models predict a future parameter as a function of past values of that same parameter.
What does the book state as the 2nd law of forecasting and how is this law explained?
Detailed forecasts are worse than aggregate forecasts.
This is because an aggregate forecast will exhibit less variability than a detailed forecast, by the concept of variability pooling.
- Higher grades + faster learning
- Never study anything twice
- 100% sure, 100% understanding
1: What is regression analysis?
2: Is it usually used for the Delphi method, causal models or time series models?
1: Basically it is a technique for fitting a function to data.
2: Causal forecasting models.
In regression analysis, what does an R square (R^2) of 0,85 mean? What about an R square of 2?
An R^2 of 0,85 means that the regression line fits the data pretty well.
An R^2 of 2 is impossible, because it can only range from 0 to 1.
The closer to 1, the better the predictive model.
For the exponential smoothing model; do higher or lower values of α make the model more responsive to changes in the process being forecast?
Same question as above, but then for values of m for the moving average model.
higher values of α
lower values of m
True or false?
Forecasts made by using the exponential smoothing or moving average model are pretty much equivalent in terms of precision and lag issues (for equivalent values of m and α).
True. Both lag behind the actual data, which is worded in point 2 (on p. 447 and p.449).
Their precision is also worded in point 2, but also on point 1 (p. 447 and p.448), which is pretty much equivalent for both models.
Furthermore, compare figure 13.14 with figure 13.15.
How can you find good values of m or α for respectively the moving average or exponential smoothing model?
Trial-and-error. However, it would be (even) better by using some optimization tools in Excel, for example.
What does it mean when f(t)=A(t)?
So what are the optimal values for MAD, MSD & BIAS?
f(t)=A(t) means that the forecast for period t is equal to the actual value in period t, so the forecast was (or is, if we can look into the future) perfect.
Since f(t)=A(t) is optimal, f(t)-A(t)=0 is also optimal. This means the optimal values for MAD, MSD & BIAS is 0.
The question on the page originate from the summary of the following study material:
- A unique study and practice tool
- Never study anything twice again
- Get the grades you hope for
- 100% sure, 100% understanding