| Week |
Lectures |
Laboratories |
Texts |
M-files |
Projects |
| 1 |
Basic MATLAB commands, matrix and vector opperations, data structures, graphics, file management, SIMULINK |
|
|
| 2
|
Matrix description of neural networks, transfer functions, classification of neural networks, selected applications
|
Basic computational algorithms, error surface, selected commands: HARDLIM, TANSIG, LOGSIG, PURELIN, ERRSURF
|
|
|
| 3
|
Fundamental model of a neuron (PERCEPTRON), network training, princple of classification
|
Perceptron learning process, selected commands: RANDS, INITP, LEARNP, TRAINP, PLOTPC, PLOTPV
|
|
|
| 4
|
Linear neural network (ADALINE), model of a neuron and network architecture, the learning proces
|
Adaptive linear network, selected commands: INITLIN, LEARNWH, TRAINWH, SOLVELIN, SIM, ADAPTWH
|
|
|
| 5
|
|
Project NN1: Adaptive rejection of signal components
|
|
|
|
| 6
|
Multi-layer neural network, model of a neuron and network architecture, Levenberg-Marquardtova method, time-series prediction
|
Basic algorithms of the learning process, optimization of network coefficients, selected commands: INITFF, SIM, LEARNBP, TRANBP
|
|
|
| 7
|
|
Project NN2: Analysis and prediction of real-time series
|
|
|
|
| 8
|
|
CASE STUDY: Adaptive Signal Processing and Adaptive Neural Networks (B. Widraw)
|
|
|
| 9
|
Functions of radial type, network model, selection of network structure
|
Network optimization, selected commands: RADBAS, SOLVERB
|
|
|
| 10
|
Associative learning methods, Hebb method, Kohonen method
|
Network traing, selected commands: LEARNH, LEARNK
|
|
|
| 11
|
Self-organizing networks, inicialization and training, self-organizing maps
|
Self-organizing maps, selected commands: NBDIST, NBGRID, COMPET
|
|
|
| 12
|
|
Project NN3: Classification of EEG signal segments
|
|
|
|
| 13
|
Modelling of neural networks in the SIMULINK environment
|
Basic SIMULINK blocks for neural networks modelling, noise reduction
|
|
|
| 14
|
|
Conclusion
|
|
|
