Browsing by Subject "programowanie genetyczne"

Now showing 1 - 5 of 5

Access status: Restricted ,
Algorithm for feature construction based on genetic programming
(Data obrony: 2020-01-24) Magiera, Franciszek
Wydział Elektrotechniki, Automatyki, Informatyki i Inżynierii Biomedycznej
Access status: Open Access ,
Algorytmy genetyczne w automatycznym generowaniu gramatyk bezkontekstowych dla potrzeb syntaktycznej interpretacji wzorców
(Data obrony: 2007) Pałka, Dariusz
Wydział Elektrotechniki, Automatyki, Informatyki i Elektroniki
The thesis presents the grammar inference system for context-free grammars based on genetic programming. This system allows for generating grammars for an unknown context-free language based on finite sets of patterns belonging to this language (i.e. positive examples) as well as patterns belonging to the complement of this language (i.e. negative examples). The thesis analyses the influence of various types of crossover operators on the process of grammar inference, as well as suggests and tests mutation operators, whose function is to increase the efficiency of the evolution process in grammatical inference. It also develops an algorithm which compensates the growth of the size of the trees, during the evolution process, which represent the production of grammars. The thesis presents the model of 'continuous' evolution, developed by the author, which is an alternative for the standard model used in genetic programming. This model is characterised by the lack of division of the evolution stages into generations and the decomposition of the set of individuals into disjunctive subsets differing from each other by parameters of evolution. The results obtained by the author indicate that 'continuous' evolution model is better suited for the inference of context-free grammars thanks to its better convergence of the process of generating grammars, and the fact that the grammars obtained are less complex.
Access status: Restricted ,
Automatyczne odkrywanie procesów biznesowych przy użyciu algorytmów programowania genetycznego
(Data obrony: 2020-10-06) Mielczarek, Łukasz
Wydział Elektrotechniki, Automatyki, Informatyki i Inżynierii Biomedycznej
Access status: Restricted ,
Nowoczesne metody określania mechanicznych właściwości gruntów
(Data obrony: 2020-09-03) Al-Richane, Fares
Wydział Wiertnictwa, Nafty i Gazu
Access status: Open Access ,
Pozyskiwanie wiedzy z dużych zbiorów danych z zastosowaniem adaptacyjnych procedur generowania zapytań
(2005-07-07) (Data obrony: 2009) Jędrzejec, Bartosz
Wydział Elektrotechniki, Automatyki, Informatyki i Elektroniki
The thesis presents a method of knowledge discovery from large databases using an evolutionary algorithm for query generation. The proposed solution is a combination of the association rules method and genetic programming. The XML-based Predictive Model Markup Language (PMML) and XQuery standard have been chosen to store and query the discovered association rule models. In the approach suggested the query is an XQuery program obtained by genetic programming. The query is modified in order to improve the quality of the selected association rules. The essential part of the thesis presents the proposed fitness function which combines a subjective measure, determined by the user, and the objective measure based on J-measure. To verify the proposed method the Java application was developed and evaluated for a number of data sets. The results of numerical experiments on the real medical, electrical power consumption and car accidents databases were presented in the final part of the thesis.