BUS338 Forecasting

This course is a broad introduction into forecasting. It employs statistical models, decision rules and programming skills to address selected business problems. The course discusses types of forecasting problems and forecastability as well as types of forecasts (point, interval, density) and their optimality under different loss functions. It introduces a number of forecasting methods ranging from simple exponential smoothing to advanced automated algorithms and presents best practices such as forecast averaging. Forecast evaluation and comparison coupled with identification of superior forecasts are also covered.

Our purpose is to acquire practical forecasting skills based on a sound understanding of statistical and decision-theoretic principles and stylized facts of business data. Teaching combines lectures, exercise sessions and independent group work on mandatory assignments.

Knowledge:

Students have advanced knowledge of

• core ideas in forecasting
• optimization criteria that underlie decision making
• randomness and its manifestations, interpretations and role in forecasting

Skills:

Students can

• formulate a forecasting problem mathematically
• apply a variety of forecasting techniques
• evaluate forecast performance
• compare alternative forecasts and select between them
• improve forecasts based on historical forecast errors and losses

General Competence:

Students

• can analyze and discuss real-world forecasting problems
• can evaluate forecasting solutions and foresee potential failures
• can use R or other forecasting software
• appreciate randomness and the human proclivity for mistaking noise for signal