1/23
Looks like no tags are added yet.
Forecasting
Predicting the future value of something using available information
2 important aspects of forecasts
Expected level of demand and Accuracy
Why are forecasts never perfect
Random variation creates some residual error even when other factors are accounted for
Forecast accuract as the time horizon increases
Accuracy generallu decrease
Qualitiative Forecasting
Uses subjective information and judgement
Quantitive Forecasting
Uses numerical data
Dephi method
An iterative process using expert opinions to reach a consensus
Forecast error
Actual value - Forecase value. Positive meand underforcasting; Negative means overforecasting
MAD
Average size of errors, ignoring whether they’re positive or negative
MSE
Gives greater weight to large errors because errors are squared
MAPE
Average error expressed as a percentage of actual values
Naive forecasting
A forecasting technique that uses a single previous value of a time series as the basis for a forecastNa
Naive casting simply put
Whatever happened last period is what we predict for the next period
Exponential Smoothing
A weighted averaging method that updates the previous forecast using a percentage of the forecast error
When alpha increases
Forecase becomes more responsive to recent changes, with less smoothingW
When alpha decreases
Forecast becomes less responsive, with more smoothing
Does higher alpha always produce a more accurate forecast?
No. It responds faster, but may react too strongly to random fluctuations
Trend adjusted exponential smoothing (TAES)
A method that cobines a smoothed forecast with a trend estimate
Why use TAES
It can respond to changed in trend, which simple exponential smoothing may lag behind.
Why should forecasts be monitored?
To determine whether forecasts are preforming satisfatorily and identify problems that may require corrective action
What is forecast bias
A systematic tendency to overforecast or underforecast
What does a positive forecast error mean?
Actual demand exceeded forecast demand; Underforecasting
What is a tracking signal used fore
Detecting forecast biasW
What can cause forecast bias
An Inadequate model, irregular variations, or random variation