Title Forecasting Techniques
Lesson Code 321-9000
Semester 8
ECTS 5
Hours (Theory) 3
Hours (Lab) 0
Faculty ICS Eng. Department

Syllabus

Time Series Data, Correlation, Time Series Analysis, Forecasting Strategies, Demanding forecasting, Basic Stochastic Models, Characteristics of Time Series, Definition of Prediction, Prediction Fields and Applications, Categories of Predictive Paths, Predictive Performance Measures, Basic statistical concepts, Statistical Methods in the Frequency Domain, Basic Statistical Analysis and prediction models, Statistical measures of accuracy in Predictions, Graphical Data Representation, Parameter Estimation, Growth Rate, Normalization Terms, classical Decomposition methods, Stationary Models, Non-stationary Models, Introduction to Spectral Analysis and Filtering, State Space Models, Multivariate Models, Confidence Space, Business Forecast Process, Mobile Intermediate Terms for Exposure, Methods of Exposure Smoothing , Seasonal Smoothing), Selection of smoothing model, Introduction to ARIMA Timeline Forecasting Models (Prediction Limits). Time Series Regression and Exploratory Data Analysis (simple linear and multiple regression), Binary Categorization (such as Support Vector Machines and Multiple Layer Perceptron) and Machine Learning applications as well as Clustering techniques (such as Neural Networks, k-Nearest Neighbours, Expectation Maximization).

Learning Outcomes

The aim of the course is to enable students to understand the basic principles of time series analysis, strategies prediction, basic Statistical Analysis and Performance measures in forecasting, Time Series Regression and Exploratory Data Analysis (simple linear and multiple regression), Binary Categorization (such as Support Vector Machines and Multiple Layer Perceptron) and Machine Learning applications as well as Clustering techniques (such as Neural Networks, k-Nearest Neighbours, Expectation Maximization). By concluding the course, students are able to:

  • analyze and adapt data in original form
  • estimate the parameters and compute the mobile average of data based on basic Statistic methodology
  • distinguish the quality of characteristics in time series data
  • apply forecasting methods analyzing and designing data required for prediction
  • develop deep knowledge in Time Series Regression and Exploratory Data Analysis
  • understand the content / role of forecasting based on basic prediction models
  • identify, describe and distinguish the main methods and prediction techniques in Binary Classification as well as clustering.
  • have comprehensive knowledge in methodology and application of forecasting techniques

 

 

Prerequisite Courses

Not required.

Basic Textbooks

  1. Μέθοδοι Προβλέψεων και Ανάλυσης Αποφάσεων, Αγιακλογλου Ν. Χρήστος, Οικονόμου Γιώργος, Εκδόσεις Μπένου, 2014.
  2. Φ. Πετρόπουλος & Β. Ασημακόπουλος, Επιχειρησιακές Προβλέψεις, Εκδόσεις Συμμετρία, 2011.
  3. Σύγχρονες Μέθοδοι Ανάλυσης Χρονολογικών Σειρών". Συγγραφέας: Δημέλη Σοφία. Εκδόσεις: Οικονομικό Πανεπιστήμιο Αθηνών. Εταιρεία Ο.Π.Α. Α.Ε.

 

Teaching and Learning Methods

Lectures, resolving exercises, Laboratory Exercises.

Activity Semester workload
Lectures 39 hours
Exercises 45 hours
Personal study 38 hours
   
Final exams 3 hours
Course total 125 hours (5 ECTS)

 

Student Performance Evaluation

Personal assignments and pair or group assignments, lab practice, regular short assessments in the form of a quiz test, final examination.

Language of Instruction and Examinations

Greek (English for Erasmus students)

Delivery Mode

Face-to-face