lect4 - Demand Forecasting II Causal Analysis Chris Caplice

Info iconThis preview shows pages 1–8. Sign up to view the full content.

View Full Document Right Arrow Icon
Demand Forecasting II Causal Analysis Chris Caplice ESD.260/15.770/1.260 Logistics Systems Sept 2006
Background image of page 1

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full DocumentRight Arrow Icon
© Chris Caplice, MIT 2 MIT Center for Transportation & Logistics – ESD.260 Agenda Forecasting Evaluation Use of Causal Models in Forecasting Approach and Methods ± Ordinary Least Squares (OLS) Regression ± Other Approaches Closing Comments on Forecasting
Background image of page 2
© Chris Caplice, MIT 3 MIT Center for Transportation & Logistics – ESD.260 Forecast Evaluation How do we determine what is a good forecast? ± Accuracy - Closeness to actual observations ± Bias - Persistent tendency to over or under predict ± Fit versus Forecast – Tradeoff between accuracy to past forecast to usefulness of predictability ± Forecast Optimality – Error is equal to the random noise distribution Combination of art and science ± Statistically – find a valid model ± Art – find a model that makes sense
Background image of page 3

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full DocumentRight Arrow Icon
© Chris Caplice, MIT 4 MIT Center for Transportation & Logistics – ESD.260 Accuracy and Bias Measures 1. Forecast Error: e t = x t - ̂ x t 2. Mean Deviation: 3. Mean Absolute Deviation 4. Mean Squared Error: 5. Root Mean Squared Error: 6. Mean Percent Error: 7. Mean Absolute Percent Error: 1 n t t e MD n = = 1 n t t e MAD n = = 1 n t t t e D MPE n = = 2 1 n t t e MSE n = = 1 n t t t e D MAPE n = = 2 1 n t t e RMSE n = =
Background image of page 4
© Chris Caplice, MIT 5 MIT Center for Transportation & Logistics – ESD.260 108.00 109.00 110.00 111.00 112.00 113.00 114.00 115.00 116.00 - 20.00 40.00 60.00 80.00 100.00 120.00 MA3 MA10 MA20 ActDemand M A 3M A 1 0M A 2 0 MD 0.05 0.21 0.35 MAD 0.56 1.07 1.41 MSE 0.47 1.67 2.71 RMSE 0.68 1.29 1.65 MAPE 0.50% 0.96% 1.27% Moving Average Forecasts
Background image of page 5

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full DocumentRight Arrow Icon
© Chris Caplice, MIT 6 MIT Center for Transportation & Logistics – ESD.260 Analysis of the Forecast Are the forecast errors ~N(0,Var(e))? ± For Moving Averages: ² What is the expected value of the errors? ² What is the variance of the errors? ± From actual observations, ² Are the observed errors ~N(0,Var(e))? ² For the MA3 data ± μ e = 0.05 ± σ e = 0.69 ± σ D = 1.478 ² Testing for Normalcy – Chi-Square, Kolmogorov-Smirnov, or other tests 0 5 10 15 20 25 ( 1. 4 0) 1 .1 (0. 7 9) 0 .4 0. 12 42 0.72 1.03 1.33 M o re error Frequency Errors (2.00) (1.50) (1.00) (0.50) - 0.50 1.00 1.50 2.00 - 2 04 06 08 01 0 0 1 2 0
Background image of page 6
© Chris Caplice, MIT 7 MIT Center for Transportation & Logistics – ESD.260 Corrective Actions to Forecasts Measures of Bias ± Cumulative Sum of Errors (C t ) ² Normalize by dividing by RMSE (U t ) ² U t should ~0 if unbiased ± Smoothed Error Tracking Signal (T t ) ² T t =z t /MAD t ² Where z t = ω e t + (1- ω )z t-1 (smoothing constant) ± Autocorrelation of forecast Errors ² Correlation between successive observations Corrective Actions ±
Background image of page 7

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full DocumentRight Arrow Icon
Image of page 8
This is the end of the preview. Sign up to access the rest of the document.

Page1 / 21

lect4 - Demand Forecasting II Causal Analysis Chris Caplice

This preview shows document pages 1 - 8. Sign up to view the full document.

View Full Document Right Arrow Icon
Ask a homework question - tutors are online