Overview of time series methods used in business administration and economics for the purpose of forecasting and forecasting evaluation with emphasis on applied forecasting:
- The decomposition of Time Series (season, cycle, trend)
- Linear models (ARIMA) including exponential smoothing
- Dynamic regression models
- Forecasting, including Judgmental forecasts
Many decisions, especially in economics and business, depend on future values of variables of interest. Hence, there is a need to be able to forecast these variables in the best possible way. For example, theoretically, the value of a stock depends on future dividends, and if you can make better forecasts of dividends, then you can price stocks better. Forecasting macroeconomic variables is essential, as they have influence on tax revenue, which is important for a government's spending opportunities. Bad forecasts may have substantial negative effects on the real economy, through bad decisions. Retail businesses may want to forecast sales volumes, which do not only depend on season, but on advertising as well, for the purpose of holding an optimal level of the product. The purpose of this course is to give the students tools to make forecast and an understanding of the modelling and forecasting of economic variables.
After completing the course the students are able to:
- understand the central ideas of time series analysis and forecasting
- decompose a time series into its components
- graphically present time series
- model a real-world time series using an appropriate time series model and use it for forecasting
- evaluate forecast performance and to identify the components of forecast errors
- use R and appropriate packages
- read scientific papers in forecasting
- develop a research question for a master thesis
- know the structure and requirements for a master thesis
Basic skills in statistical inference.
Requirements for course approval
The assessment for this course will not be changed in the spring of 2020.
The final grade will be based on two major projects that accounts for 50 % each. The students can work in groups (1-2 students in each group).
Students will be working on the first project in week 10-12, and they will be working on the second project in week 12-14.
Students can retake each project individually.
We will make use of the statistical packages in R.
Forecasting: Principles and Practice, Rob J. Hyndman and George Athanasopoulos, 2nd edition (available at https://otexts.org/fpp2/).
- ECTS Credits
- Teaching language
Spring. Offered Spring 2020.
Adjunct Professor Johan Lyhagen, Department of Business and Management Science.