Academic year2016/17
CoursePHYSICS
Code199BB
Credits6
PeriodSemester 1
LanguageItalian
Modules | Area | Type | Hours | Teacher(s) |
ELABORAZIONE DEI SEGNALI | FIS/01 | LEZIONI | 36 | |
Programma non disponibile nella lingua selezionata
Knowledge
Students are expected to acquire:
knowledge to extract signal from noise applying:
- optimum/adaptive filtering (regressive model)
- principal component / total least squares modelling
- independent component model
some knowledge of pattern recognition / classification:
- elements of Bayesian decision theory
- discriminant functions
- clustering
Assessment criteria of knowledge
The student's ability to explain correctly the main topics presented during the course at the board will be assessed.Methods:
Further information:
Final oral exam 100%
Teaching methods
Delivery: face to face
Learning activities:
Attendance: Mandatory
Teaching methods:
Syllabus
Optimum filtering: Wiener filter; adaptive filters (LMS and RLS algorithm).
Autoregressive model, parametric spectrum estimation.
Total least squares model, singular value decomposition, signal and noise subspaces.
Independent Component Analysis: maximization of non gaussianity, maximization of likelihood, minimization of mutual information.
Pattern recognition/classification: feature selection (principal component analysis, generalized variance ratio);
Elements of Bayesian decision theory; maximum likelihood and Bayesian estimation; linear and quadratic discriminant functions. Classifier training and validation.
Clustering: k-means; hierarchical clustering.
Bibliography
Recommended reading: Professor's lecture notes.
Additional reading includes:
- S. Haykin. Adaptive Filter Theory. Prentice Hall.
- A. Hyvarinen, J. Karhunen, E. Oja. Independent Component Analysis. Wiley-Interscience Publication.
- R. O. Duda, P. E. Hart and D. G. Stork. Pattern Classification. Wiley Interscience.
Further bibliography:
S.T. Alexander. Adaptive Signal Processing, Theory and Applications. Springer-Verlag.
B. Widrow, S.D. Stearns. Adaptive Signal Processing. Prentice Hall.
S.L.Marple. Digital spectral estimation. Prentice-Hall.
S. Van Huffel, J. Vanderwalle. The Total Least Squares Problem: Computational Aspects and Analysis. SIAM, Philadelphia, PA, 1991.
G.H. Golub, C. Van Loan. Matrix Computations. The Johns Hopkins University Press.
Updated: 14/11/2016 17:27