Combination of lectures, exercises, guided lab activities (MATLAB)
• Data Analysis and Data visualization (wk 1): Types of data. Analysis as modeling. Statistical data analysis. The do’s and don’t’s of data visualization • Probability density estimates (wk 2-4): Unsupervised learning. Gaussian model. Principal Component Analysis, Factor Analysis, Independent Component Analysis, Cluster analysis and the EM algorithm. • Pattern analysis and decision theory; classifiers (wk 5-6): Bayesian decision theory. Bayes classifiers. Logistic classifiers. Performance of a classifier: ROC curve. • Multilayer neural networks (wk 7): Generalized linear models. Perceptrons. Multilayer neural networks and the backpropagation algorithm. Design of a neural network model • Graphical Models (wk 8-10) Static data. A general framework for data modeling. The EM algorithm. Regression, decision theory, factor analysis as graphical models. Dynamic graphical models (overview): Temporal data (signals). Discrete vs continuous signals. Hidden Markov Models, Linear dynamical systems as dynamic graphical models • Generalization (wk 11-12) : Regularization theory, Vector Quantization. Support Vector Machines.
Ricevimento: on demand, by e-mail to: vittorio.sanguineti@unige.it or mobile phone at: 3292104393. Teacher office: via All’Opera Pia 13, building E, fourth floor. Office direct phone number: 010-3536487
MAURO GIACOMINI (President)
VITTORIO SANGUINETI (President)
MARCO MASSIMO FATO
ANALYSIS OF BIOMEDICAL DATA AND SIGNALS
• Written examination (weight 50%) • Project work (individuals or couples, weight 50%): Solution of a real problem of biomedical data analysis/processing, chosen from a list of proposed projects; Development of software for calculation/analysis/processing; Final report presenting the results, structured as a scientific article
