Forecasting Techniques
Objectives
The course aims to provide students with knowledge on the application of forecasting methodologies, models and techniques, mainly to support management decision making. This course has a strong practical component of formulation, modeling and solving problems in laboratory.
At the end the course it is intended that students have acquired the skills necessary to understand how the application of forecasting techniques contributes positively on the effectiveness and efficiency on management of both supply chain and organization; have an overview of a comprehensive set of forecasting techniques and a perception of the strengths and weaknesses of each technique, and also develop skills that allow them to select the forecasting models more suited to the specificities of both data modeling and the goal that should be achieved by forecast.
Additionally it is intended that students will be able to identify relevant aspects with regard to the forecast process, data considerations, model selection, and forecast implementation in large scale problems. It is intended also that students are able to develop a critical sense regarding the importance of demand forecasting in supply chain management, production management, marketing and finance.
General characterization
Code
7201
Credits
6.0
Responsible teacher
Ana Paula Ferreira Barroso, António Carlos Bárbara Grilo
Hours
Weekly - 4
Total - 61
Teaching language
Português
Prerequisites
Available soon
Bibliography
Hanke J. E. e Wichern D. W. (2009) Business Forecasting. Pearson International Edition.
Wilson J.H., Keating B. e Galt J. (2009) Business Forecasting with ForecastX. McGraw Hill.
Hoshmand A. R. (2010) Business Forecasting. A practical approach. Routledge, Taylor & Francis Group.
DeLurgio S. A. (1998) Forecasting Principles and Applications. Irwin McGraw-Hill.
Box G.E.P., Jenkins G.M. e Reinsel G. C. (1994) Time Series Analysis, Forecasting and Control, 3th ed., Englewood Cliffs, Prentice-Hall.
Teaching method
In lectures the concepts, methods and models are discussed. Oral questions are frequently made for prerequisite control, knowledge assessment and stimulate students’ participation.
In laboratory sessions the experimental method is adopted. Active methods are used. Students are challenged with multifaceted problems which should be solved in team. Also, case studies are analyzed and discussed in class.
Evaluation method
The assessment will be carried out remotely. It includes 3 components with the respective weighting in the final grade:
a) Theoretical-Practical Assessment (40%) - Individual
2 tests (T1 and T2). The average of the 2 tests must have a minimum score of 9.5 (out of 20) for approval in UC
b) Laboratory or Project Assessment (45%) - In group
3 group projects (Trbs)
c) Summative Assessment (15%) - Individual
5 quizzes throughout the semester (mT)
Final grade = 0.15 mT + 0.20 T1 + 0.20 T2 + 0.45 Trbs
The score for each component of the assessment will be rounded to two decimal places.
T1: 2 November ; T2: 14 December
Students are excluded from the final exam if they fail the quizzes.
The final grade must be at least 9.50 (out of 20). If it is less than 9.50 values, the student will have to take the final exam (Exame de Recurso), which replaces only T1 and T2.
The student must make the Grade Defense if the final grade is higher than 17 values. The Grade Defense is carried out through an Oral Exam to be scheduled with the student between 3 and 7 days after the final grade is released.
Subject matter
- Planning and forecasting. Types of forecasting. Statistical fundamentals for forecasting. Quantitative and qualitative forecasting
- Exploring time series data patterns. Adjusting outliers in time series with and without seasonal pattern.
- Fitting versus forecasting: absolute and relative measures of error. Autocorrelation and ACF (k)
- Univariate methods to model time series without trend or seasonality: simple smoothing methods: simple moving averages, weighted moving averages, exponential smoothing
- Univariate methods to model time series with trend (no seasonality). Linear regression models. Estimating trends with differences. Brown’s model. Holt’s model
- Univariate methods to model time series with trend and seasonality. Holt-Winters’ model. Multiplicative decomposition method. Additive decomposition method. Decomposition using regression models
- Univariate ARIMA models. ARIMA applications