The course intends to provide the basic notions for data and signal analysis with emphasis on application to biology and medicine.
The course provides students with the essential tools and operational skills for quantitative analysis of data and signals of interest for medicine and biology, on a probabilistic perspective
By the end of the course the students will be able to:
1. Design and apply methods of analysis and modeling of data - including temporal data (signals) - of interest for medicine and biology
2. Identify the correct approach (model selection, model identification, data visualization) for a specific data analysis problem
3. Use MATLAB to display and model biomedical data and signals
There are no formal prerequisites, but the course requires solid foundations in calculus and linear algebra.
The course is organized as a combination of lectures and practical activities.
Lectures will focus on theory and methods for data analysis.
Practical activities will focus on application of theory to real data analysis problems in the context of bioengineering.
A. Data Analysis and Data Display. Data types. Descriptive statistics. Analysis as modeling. Statistical data analysis. Regression. Visual display of information.
B. Probability density estimates. Unsupervised learning. Gaussian model. Principal Component Analysis, Factor Analysis, Independent Component Analysis, Cluster analysis and the EM algorithm. Graphical models. Regression and factor analysis as graphical models.
C. Pattern analysis and decision theory. Bayesian decision theory. Bayes classifiers. Logistic classifiers. Performance of a classifier: ROC curve. Generalized linear models. Introduction to neural networks. Model generalization and bias-variance trade-off.
D. Model selection. Statistical Inference. Hypothesis testing. General Linear models and the analysis of variance. Mixed-effects models. Bayesian model selection.
E. Dynamic Models. Temporal data (signals). Hidden Markov Models, Linear dynamical systems. Kalman filters.
There is no course textbook; the course material draws from a variety of sources. Recommended references:
1. Murphy, KP (2022) Probabilistic Machine Learning: An Introduction. MIT Press.
2. Bishop, CM (2006) Pattern Recognition and Machine Learning. Springer
3. Barber D (2006) Machine Learning A Probabilistic Approach.
Ricevimento: VITTORIO SANGUINETI. Appointment: Tel. 0103356487 or vittorio.sanguineti@unige.it
Ricevimento: For appointment contact: martina.brofiga@unige.it
VITTORIO SANGUINETI (President)
MARTINA BROFIGA (President Substitute)
https://corsi.unige.it/11159/p/studenti-orario
The timetable for this course is available here: EasyAcademy
Written exam (weight 50%)
Project work (individuals or couples, weight 50%)
Project work will be assessed in terms of:
1) Documentation (correctess, clarity, synthesis, terminology)
2) Implementation (code structure and organization, efficiency)
3) Data Visualization (technical quality of figures, adequacy of display technique, efficacy, clarity)
4) Bonus (max 2 pts) if the report provides additional analysis (in addition to those required). Bonus is only awarded if maximum score is obtained in the other three criteria.
Ask the teacher for other information not included here
