Course overview
The aim of this course is to impart knowledge, skills, and methods that can be used to commission and conduct forecasting projects that produce useful forecasts for diverse management decisions, and to review forecasting methods used by others to determine whether or not their forecasts are valid. Forecasts can be of decisions, outcomes, quantities, and time series for forecasting problems including, but not limited to, those involved with demand, store location, hiring, population numbers, conflict situations, and governance and regulation.
Introduction: Forecasting's domain, role of unaided judgment, problem structuring and method selection
Meetings and markets: Nominal groups, estimate-talk-estimate, and prediction markets
Surveys: Intentions and expectations, conjoint analysis, judgmental bootstrapping, and Delphi
Analogs: Structured and quantitative analogies
Experiments: Experimentation, conjoint analysis, and simulated interactions
Extrapolation: Naïve models, exponential smoothing, damped trend
Causal models: Regression analysis, segmentation, and the index method
Combining and adjustment: Combining forecasts, forecasters, and methods, and adjusting for omitted information
Validation and checklists: Assessing the accuracy of forecasts from alternative methods, and Simple Forecasting and Golden Rule checklists
Presentation: Prediction intervals, scenarios, and persuasive reports and talks.
Course learning outcomes
- Be alert to the broad relevance of forecasting and understand the role of forecasting as the basis of good decisions
- Structure decision making problems to make best use of evidence-based forecasting methods
- Identify and apply appropriate simple forecasting methods so as to provide forecasts for decision makers
- Present forecasts to decision makers in a useful and ethical way
- Critically assess the validity and reliability of forecasting procedures proposed and used by others