Data modelling

Lecturer: Janez Povh

Syllabus outline:

• Introduction to data modelling: what is data model, how to compute and validate data model;
• Sources of data: sensor data; data repositories; open data; data warehouses,…;
• Supervised learning: regression, classification (neural networks, logistic regression, support vector machines); analysis of supervised learning models (cross-validation, confusion matrix, precision, accuracy, recall); bootstrapping
• Unsupervised learning: clustering, principal component analysis; evaluating the unsupervised models (purity, normalized mutual information, Rand index,…)
• Data Mining methods: association rules, decision trees
• Recommendation systems: data model for a recommendation systems, content-based recommendations;
• Deep learning: image and speach recogniiton
• Data modelling with state of the art open source software: R, WEKA, Orange

Objectives and competences:

• the use of methodological tools, ie. implementation, coordination and organization of research, application of various quantitative research methods and techniques
• the use and combining the knowledge from different disciplines
• the ability to use information and communications technologies and data analytic tools in engineering
• ability to collect, store, analyse and interpret big data
Subject-specific competences:
• ability of collecting data and performing and sustainable management of data;
• ability of creating and validating advanced data models;
• mastering supervised and unsupervised statistical learning methods;
• mastering the key data mining methods;
• mastering at least one state-of-the- art tool for statistical modelling (R, Weka, Orange)

Intended learning outcomes:

The student will:
• understand the importance and potentials of data modelling;
• master the key statistical methods underlying the data modelling;
• learn how to use state-of-the-art software tools to perform advanced data modelling (R, Weka, Orange)