Information Modeling & Analysis
We live in a world where big amounts of data are generated by multiple sources. Among others, software development originates a constant stream of data about process, code and executions. However, data is useless unless it is properly modelled and analyzed. This course will teach students how to turn data into information, i.e., processed and organized data with meaning, by automated identification of rules, patterns and regularities. This involves:
- getting familiar with machine learning algorithms, their assumptions and their limitations;
- learning how to select the training corpus and how to evaluate the outcome of training;
- applying machine learning algorithms to a set of case studies, taken from recent research in Software Engineering.
The course will cover unsupervised learning (e.g., clustering, feature maps), supervised learning (e.g., classifiers, neural networks) and concept mining (association rules, latent topics).