Κατεύθυνση: Major in Communications & Networks

Data Warehouses and Data Mining

Learning Outcomes

The students upon the successful completion of the course will be able:

  • to evaluate the quality of the data to be analyzed and apply the appropriate data pre-processing techniques,
  • to select the appropriate data mining technique based on requirements and data type,
  • to design and develop data warehouses,
  • to use the appropriate data mining techniques and tools to extract knowledge from data collections,
  • to evaluate the quality of data mining results.

Course Contents

  • Introduction to the fundamental data mining concepts and techniques: main steps of knowledge and data discovery, requirements of developing data mining approaches.
  • Data pre-processing: data cleaning, transformation, dimensionality reduction.
  • Data warehouses: multidimensional models, architecture, implementation of data warehouses, OLAP.
  • Clustering: partitional, hierarchical, density-based, grid-based, spectral clustering, clustering applications.
  • Classification: Bayesian classifiers, decision trees, k-nearest neighbors.
  • Association rules: Apriori, representative association rules.
  • Quality assessment in data mining: evaluation of classification models, association rules interestingness measures, cluster validity.
  • Web mining: link analysis, text mining, web search, PageRank.

Recommended Readings

  • Han J. & Kamber M. (2006): Data Mining: Concepts and Techniques, 2nd Edition, Morgan Kaufmann.
  • Chakrabarti S. (2002): Mining the Web, Discovering Knowledge from Hypertext Data, Morgan Kaufman Publishers.

Development of Telecommunication Systems

Learning Outcomes

The aim of this course is to familiarize students with best practices for the design and implementation of modern innovative telecom systems and applications. The main topics discussed include systematic multi-criteria requirement analysis, transformation of user requirements to architectural specifications and system functionality, design optimization for high performance and QoS, installation and deployment procedures, as well as system validation against initial requirements and specifications. The presented methodologies derive analytically from real case studies of innovative telecommunication systems and applications with extremely high added value and impact (e.g. the European Space Agency SATWAYS project in the field of Air Traffic Control). Students will further gain significant practical experience in the design and implementation of innovative telecommunication systems through exercise solving and development of small projects. Last but not least, the course presents an in-depth commercial feasibility analysis of the developed systems.

At the end of the course, students will be equipped with advanced expert and analytical knowledge for the consistent design, development and validation of innovative embedded telecommunication systems (see Course Content) meeting strict quality requirements regarding the provided telecommunication services. The obtained knowledge will allow the critical and analytical deepening as well as performing innovative research in the broad scientific domain of embedded telecommunication systems and applications.

Students will be capable of:

  • transforming user requirements to telecommunication systems architectural specifications and functionality.
  • specifying prototype embedded telecommunication systems, including their architectural layering (hardware, embedded software, software), according to specific QoS requirements for the provided services.
  • designing and specifying modern voice communication systems for next generation networks.
  • designing QoS models involving the end-to-end telecommunication transmission path from service provisioning to the end user, as well as extracting performance requirements and specifications for segments of the end-to-end transmission path.
  • optimizing the design of telecommunication systems for high performance and QoS.
  • explaining the details of the E-model adaptation and simplification of the discrete noise and advantage parameters in the demanding field of radio voice communications. The E-model (ITU-T Rec. G.107) is a transmission planning tool that provides a prediction of the expected voice quality, as perceived by a typical telephone user, for a complete end-to-end (i.e. mouth-to-ear) telephone connection under conversational conditions.
  • exploiting and solving the adapted voice QoS model for extracting system performance design parameters (e.g. for selecting appropriate voice codecs assuming a limited end-to-end delay and the constant values of delay parameters along the telecommunication transmission path segments).
  • compiling detailed plans for technical validation of system deliverables against the initial user requirements and system specifications.
  • analyzing the satellite link budget for the provisioning of telecommunication services and extracting requirements regarding the underlying telecommunication system.
  • analyzing the installation, deployment and operational details of the developed systems.
  • compiling workplans, work packages, time schedules, lists of technical deliverables and reports, critical paths, milestones and sample budgets involving R&D projects that design and develop innovative embedded telecommunication systems.
  • performing indicative techno-economic studies and commercial feasibility analysis regarding innovative embedded telecommunication systems, including a SWOT analysis (strengths, weaknesses, opportunities, threats).

Course Contents

  • Exemplary Innovative telecommunication systems and applications with high added value and impact.
  • Systematic multi-criteria requirement analysis.
  • Transformation of user requirements to architectural specifications and system functionality.
  • Design optimization for high performance and QoS.
  • Adaptation and validation of the E-model in the radio communications use case, including satellite transmission. The E-model (ITU-T Rec. G.107) is a transmission planning tool that provides a prediction of the expected voice quality, as perceived by a typical telephone user, for a complete end-to-end (i.e. mouth-to-ear) telephone connection under conversational conditions.
  • Exploitation of the voice QoS model for extracting system performance design parameters (e.g. for selecting appropriate voice codecs assuming a limited end-to-end delay and the constant values of delay parameters along the telecommunication transmission path segments).
  • Voice communications based on the NGN (Next Generation Network) standard.
  • Installation procedures and relevant issues.
  • System validation against initial requirements and specifications.
  • Commercial feasibility analysis.
  • Exercises and programming projects.

Recommended Readings

  • Apostolacos S. & Meliones Α. (2014): Satellite IP Radio Communications in Air Traffic Control: Design, Implementation and Evaluation of Telecommunication Systems (in Greek).
  • S. Apostolacos, A. Meliones, S. Badessi, G. Stassinopoulos, “Adaptation of the E-model for satellite internet protocol radio calls in Air Traffic Control”, IEEE Transactions on Aerospace and Electronic Systems, 50(1), January 2015.
  • S. Apostolacos, M. Manousos, A. Meliones, D. Kavadas, G. Lykakis, A. Manousarides, M. Kardaris, K. Simeakis, Design and Implementation of a Solution for the Provisioning of Converged Remote Tower and Facility Management Services over Satellite IP for Greek Heliports, IEEE Communications Magazine, 46(8), August 2008.

Information Retrieval

Learning Outcomes

The aim of this course is learning fundamental concepts of information retrieval systems. The course’s contents cover all stages of system design and implementation for collection, indexing and searching of text documents, as well as evaluation methods. In addition, recent trends in information retrieval are also covered, for example information retrieval from the WWW.

Upon successful completion of the course, the students will be in position:

  • to know representation models for text documents.
  • to use techniques for indexing, compression, retrieval and scoring of documents.
  • to develop applications that manage large volumes of text.
  • to build the functionality of a search engine.
  • to apply machine learning techniques for text classification.

Course Contents

  • Introduction and basic IR concepts
  • System architecture of IR systems
  • Dictionaries and inverted indexes
  • Construction and compression of dictionaries
  • Information retrieval models (boolean model, vector space model, probability models)
  • Scoring and ranking documents
  • Language models
  • Information retrieval from XML documents
  • Basic concepts of information retrieval from the WWW
  • Web crawling and indexing
  • Text classification with machine learning techniques, support vector machines, algorithms for text classification

Recommended Readings

  • Christopher D. Manning, Prabhakar Raghavan and Hinrich Schütze, Introduction to Information Retrieval, Cambridge University Press. 2008.
  • Ricardo A. Baeza-Yates and Berthier Ribeiro-Neto. 1999. Modern Information Retrieval. Addison-Wesley Longman Publishing Co., Inc., Boston, MA, USA.

Pattern Recognition

Learning Outcomes

Pattern recognition is the scientific field that deals with the assignment of a label to a given input value. An example of pattern recognition is classification, which attempts to assign each input value to one of a given set of classes. The course aims to cover the most popular in the literature techniques for pattern recognition, as they are typically employed in a number of practical applications, such as speech and audio recognition, image and video analysis, biometrics and bioinformatics. The course covers the most commonly used classification algorithms, feature selection techniques, data transformation methods, and data clustering.

Students, upon successful completion of the course, will be able to:

A) Understand the key standards recognition methodologies

B) Analyze problems in various areas of application, such as voice and audio recognition, image and video analysis, biometrics and bioinformatics.

C) Choose the best classifiers, feature selection methods, data transformations, and clustering.

D) Evaluate standard pattern recognition systems

 

Course Contents

  • Introduction to Pattern recognition systems
  • Parametric estimation of probability density function (maximum Likelihood estimation, maximum a posteriori
  • Bayesian classifiers and Bayesian Networks
  • k-nearest neighbor
  • Non parametric estimation of probability density function (Parzen windows)
  • Linear classifiers, non linear classifiers. Perceptron algorithm. Multilayer neural networks, Deep Learning
  • Unsupervised Pattern recognition – Clustering
  • Feature generation: contour representation and contour tracing, chain code, polygon, signatures, linear transforms, Fourier Transform, regional features, image recognition, bias and variance, texture
  • Feature Selection and Kernels
  • Pattern recognition tools

Recommended Readings

  • Sergios Theodoridis and Konstantinos Koutroumbas. 2008. Pattern Recognition, Fourth Edition (4th ed.). Academic Press

Algorithms for Electronic Markets

Learning Outcomes

The course’s material includes the theory and practice that pertain to the design of economic mechanisms for automated trade exchanges, in modern digital platforms (auction websites, services provision and products retail websites, internet advertisement platforms). In particular, the course concerns the modern algorithmic techniques that facilitate the digital implementation of electronic markets.

Upon successful completion of the course, the students will be in position:

  • to understand the economic and algorithmic background that underlies the functionality of electronic markets.
  • to design electronic trade exchanges platforms, by choosing the appropriate economic mechanisms and the relevant algorithmic implementations.
  • to assess and evaluate the performance of economic mechanisms and their algorithmic implementations, relative to a given electronic market and its particulars.
  • to design, implement and evaluate automated pricing mechanisms.

Course Contents

  • Introduction to Game Theory: Strategies, Utility Functions
  • Strategic Games and Nash Equilibrium
  • Efficiency of Equilibria
  • Oligopoly Models
  • Auctions: First-Price, Second-Price, Multi-Unit Formats
  • Algorithmic Mechanism Design
  •  Sponsored Search Auctions
  • Combinatorial Auctions
  • Principles and Methods of Pricing
  • Prediction Techniques
  • Online Auctions

Recommended Readings

  • N. Nisan, T. Roughgarden, E. Tardos, V. Vazirani. Algorithmic Game Theory. Cambridge University Press, 2006.
  • T. Roughgarden. Twenty Lectures on Algorithmic Game Theory. Cambridge University Press, 2016.
  • M. J. Osborne. An Introduction to Game Theory. Oxford University Press, 2009.
  • R. Gibbons. A Primer in Game Theory. Financial Times / Prentice Hall, 1992.