Special topology of an artificial neural network used as an approximator and classifier
Provider: Univerzita Pardubice
Programme: Studentská grantová soutěž
Implementation period: 01.01.16 - 31.12.16
Workplace:
Fakulta elektrotechniky a informatiky - Katedra řízení procesů
Investigator: Doležel Petr
Description:
Artificial neural network is a robust tool for dynamic systems modelling as well as for multidimensional data classification. However, generally provided topologies are not suitable for further analysis. An actual work of the project participants focuses on finding the suitable topology of the network, so that the network can be transformed to the system better suitable for mathematical analysis. The main usage of the approach should be in predictive control, signal analysis, image processing, etc.
Artificial neural network is a robust tool for dynamic systems modelling as well as for multidimensional data classification. However, generally provided topologies are not suitable for further analysis. An actual work of the project participants focuses on finding the suitable topology of the network, so that the network can be transformed to the system better suitable for mathematical analysis. The main usage of the approach should be in predictive control, signal analysis, image processing, etc.
