Big Data Mining

The information circulating on the web and social networks is increasingly multimedia in nature. The possibility to understand the content and to search for multimedia documents on a large scale, expecially in the absence of textual descriptions, has become a strategic tool. The module aims to present tools for analyzing and extracting information from multimedia data, in order to search them in huge databases. This module is based on hands-on work.

This module introduces the main methods of analysis and mining of opinions and personal evaluations for users based on Big Data generated on the web or other sources. Emphasis will be put on text mining method applied to text originated on social media. Lessons will be supported by case studies developed in the SoBigData.eu lab.

This module introduces the main methods of analysis and mining of opinions and personal evaluations for users based on Big Data generated on the web or other sources. Emphasis will be put on text mining method applied to text originated on social media. Lessons will be supported by case studies developed in the SoBigData.eu lab.

The purpose of the course is to introduce the main techniques in data mining and machine learning (including deep learning approaches) for the analysis of temporal data, in particular for time series and spatio-temporal data related to human mobility. The presentation will be supported by several case studies developed with the SoBigData.eu Laboratory.
 

The course will first introduce the fundamentals of artificial neural networks and, then, it will provide an overview of the main techniques and models of the deep learning field. Specific focus will be placed on detailing neural models that are useful for addressing predictive tasks on vectorial, sequential and image data, and to generative deep learning, including variational and adversarial learning.

The module provides an introduction to base concepts of data mining and knowledge extraction process, introducing analytical models and algorithms for clustering, classification and pattern discovery, also referring Big Data sources.

The module addresses practical aspects of machine learning and neural networks. It presents and reviews the main technological solutions to solve two machine learning problems: classification and regression. The course covers several crucial aspects to take into account when developing machine/deep learning solutions: i) what is the best solution to adopt for a given problem? ii) how to evaluate a machine learning model? iii) how to optimize it?

The module provides an introduction to base concepts of data mining and knowledge extraction process, introducing analytical models and algorithms for clustering, classification and pattern discovery, also referring Big Data sources.

In this course we start from the basic notions of graph theory and self-similar phenomena in order to correctly analyse large socio-economic networks. From this analysis we then proceed in the description of the modelling for various classes of phenomena and to the correct definition of benchmarks through an approach inspired by classical statistical physics.

The course presents the main elements for understanding financial markets, their structure, and technological infrastructure. Specifically, the course provides a background on basic empirical modeling of financial time series, from low to ultrahigh frequency, identifying the key data science aspects including data storage, latency, high dimensional inference, etc. It also covers semantic analysis of texts from news feed and social networks for financial forecasting.