| |
|
Knowledge Discovery in Spatio-Temporal Data
Pešek, Martin ; Bartík, Vladimír (referee) ; Zendulka, Jaroslav (advisor)
This thesis deals with knowledge discovery in spatio-temporal data, which is currently a rapidly evolving area of research in information technology. First, it describes the general principles of knowledge discovery, then, after a brief introduction to mining in the temporal and spatial data, it focuses on the overview and description of existing methods for mining in spatio-temporal data. It focuses, in particular, on moving objects data in the form of trajectories with an emphasis on the methods for trajectory outlier detection. The next part of the thesis deals with the process of implementation of the trajectory outlier detection algorithm called TOP-EYE. In order to testing, validation and possibility of using this algorithm is designed and implemented an application for trajectory outlier detection. The algorithm is experimentally evaluated on two different data sets.
|
|
Data mining analysis of chemical bonds in alloys
Nechutová, Vendula ; Šeda, Miloš (referee) ; Roupec, Jan (advisor)
The thesis deals with aplication of data mining methods for the analysis of two Ni3Si supercells, one with a stable grain boundary and the second one with unstable grain boundary. DOS and COHP curves are examined using selected curve matching methods. The surroundings of the individual atoms are examined by the Voronoi diagram. This information was used to reveal the differences in binding between stable and unstable supercell.
|
|
Data mining analysis of chemical bonds in alloys
Nechutová, Vendula ; Šeda, Miloš (referee) ; Roupec, Jan (advisor)
The thesis deals with aplication of data mining methods for the analysis of two Ni3Si supercells, one with a stable grain boundary and the second one with unstable grain boundary. DOS and COHP curves are examined using selected curve matching methods. The surroundings of the individual atoms are examined by the Voronoi diagram. This information was used to reveal the differences in binding between stable and unstable supercell.
|
| |
|
Knowledge Discovery in Spatio-Temporal Data
Pešek, Martin ; Bartík, Vladimír (referee) ; Zendulka, Jaroslav (advisor)
This thesis deals with knowledge discovery in spatio-temporal data, which is currently a rapidly evolving area of research in information technology. First, it describes the general principles of knowledge discovery, then, after a brief introduction to mining in the temporal and spatial data, it focuses on the overview and description of existing methods for mining in spatio-temporal data. It focuses, in particular, on moving objects data in the form of trajectories with an emphasis on the methods for trajectory outlier detection. The next part of the thesis deals with the process of implementation of the trajectory outlier detection algorithm called TOP-EYE. In order to testing, validation and possibility of using this algorithm is designed and implemented an application for trajectory outlier detection. The algorithm is experimentally evaluated on two different data sets.
|
| |