{"id":{"repo_id":"vilnius","oai_identifier":"oai:vu.lt:elaba:308337942"},"canonical_url":"https://search.ndltd.org/etd/vilnius/oai:vu.lt:elaba:308337942","repository":{"repo_id":"vilnius","name":"Vilnius University","base_url":"https://epublications.vu.lt/oai"},"display":{"title":"Dirbtinių neuroninių tinklų apmokymas su triukšmu /","abstract":"This thesis examines the impact of noise in training data on the performance of an artificial neural network and the $k$-nearest neighbors method in regression tasks. For the practical experiments, artificial datasets were generated using the Sphere, Rosenbrock, and Rastrigin benchmark functions. Gaussian noise was applied to introduce noise into the data. Practical experiments were conducted using both methods on clean and noisy datasets of different sizes. The performance of the methods was evaluated using the mean absolute error and standard deviation. The results showed that the artificial neural network utilized increasing amounts of data more effectively when approximating the Sphere and Rosenbrock functions. In contrast, performance improvement on the Rastrigin function was limited by the selected architecture with a sigmoid activation function in the hidden layer. Regarding the $k$-nearest neighbors method, a consistent improvement in performance was observed as the dataset size increased. This characteristic allowed the method to achieve better results for the Rastrigin function, which was more difficult for the artificial neural network to approximate. The study also found that the effect of added noise was greater for functions whose structure the artificial neural network learned more accurately. The lower impact of noise in the case of the Rastrigin function was associated with the model’s limited ability to represent its complex structure.","abstract_html":"This thesis examines the impact of noise in training data on the performance of an artificial neural network and the $k$-nearest neighbors method in regression tasks. For the practical experiments, artificial datasets were generated using the Sphere, Rosenbrock, and Rastrigin benchmark functions. Gaussian noise was applied to introduce noise into the data. Practical experiments were conducted using both methods on clean and noisy datasets of different sizes. The performance of the methods was evaluated using the mean absolute error and standard deviation. The results showed that the artificial neural network utilized increasing amounts of data more effectively when approximating the Sphere and Rosenbrock functions. In contrast, performance improvement on the Rastrigin function was limited by the selected architecture with a sigmoid activation function in the hidden layer. Regarding the $k$-nearest neighbors method, a consistent improvement in performance was observed as the dataset size increased. This characteristic allowed the method to achieve better results for the Rastrigin function, which was more difficult for the artificial neural network to approximate. The study also found that the effect of added noise was greater for functions whose structure the artificial neural network learned more accurately. The lower impact of noise in the case of the Rastrigin function was associated with the model’s limited ability to represent its complex structure.","abstract_has_math":true,"creators":["Rinkevičiūtė, Gabrielė,"],"institution":"Institutional Repository of Vilnius University","degree_name":null,"degree_level":null,"degree_discipline":null,"degree_department":null,"school":null,"contributors":[],"advisors":[],"committee_chairs":[],"committee_members":[],"year":2026,"date_issued":"2026","date_published":"2026","updated_at":"2026-11-19T10:12:59Z","subjects":[],"languages":["lit"],"rights":["info:eu-repo/semantics/openAccess"],"rights_urls":[],"identifier_entries":[],"person_identifier_entries":[]},"links":{"outbound_url":"https://repository.vu.lt/VU:ELABAETD308337942&prefLang=en_US","outbound_label":"Repository record","outbound_source":"dc:identifier","record_url":"https://repository.vu.lt/VU:ELABAETD308337942&prefLang=en_US","record_label":"Repository record","record_kind":"record","record_source":"dc:identifier","download_url":null,"download_label":null,"download_source":null,"access_status":"open","embargo_until":null,"access_note":null},"source_record":{"url":"https://epublications.vu.lt/oai?verb=GetRecord&metadataPrefix=oai_dc&identifier=oai%3Avu.lt%3Aelaba%3A308337942","prefix":"oai_dc"},"metadata_groups":[{"id":"people","label":"People","entries":[{"key":"dc:creator","label":"Author","values":["Rinkevičiūtė, Gabrielė,"]}]},{"id":"academic_context","label":"Academic Context","entries":[{"key":"dc:date","label":"Dc Date","values":["2026"]},{"key":"dc:publisher","label":"Institution","values":["Institutional Repository of Vilnius University"]},{"key":"dc:relation","label":"Dc Relation","values":["https://epublications.vu.lt/object/elaba:308337942/308337942.pdf"]},{"key":"dc:type","label":"Dc Type","values":["info:eu-repo/semantics/bachelorThesis"]}]},{"id":"language_rights","label":"Language and Rights","entries":[{"key":"dc:language","label":"Dc Language","values":["lit"]},{"key":"dc:rights","label":"Dc Rights","values":["info:eu-repo/semantics/openAccess"]}]},{"id":"identifiers","label":"Identifiers","entries":[{"key":"dc:identifier","label":"Identifier","values":["https://repository.vu.lt/VU:ELABAETD308337942&prefLang=en_US"]}]},{"id":"additional","label":"Additional Metadata","entries":[{"key":"dc:description","label":"Description","values":["This thesis examines the impact of noise in training data on the performance of an artificial neural network and the $k$-nearest neighbors method in regression tasks. For the practical experiments, artificial datasets were generated using the Sphere, Rosenbrock, and Rastrigin benchmark functions. Gaussian noise was applied to introduce noise into the data. Practical experiments were conducted using both methods on clean and noisy datasets of different sizes. The performance of the methods was evaluated using the mean absolute error and standard deviation. The results showed that the artificial neural network utilized increasing amounts of data more effectively when approximating the Sphere and Rosenbrock functions. In contrast, performance improvement on the Rastrigin function was limited by the selected architecture with a sigmoid activation function in the hidden layer. Regarding the $k$-nearest neighbors method, a consistent improvement in performance was observed as the dataset size increased. This characteristic allowed the method to achieve better results for the Rastrigin function, which was more difficult for the artificial neural network to approximate. The study also found that the effect of added noise was greater for functions whose structure the artificial neural network learned more accurately. The lower impact of noise in the case of the Rastrigin function was associated with the model’s limited ability to represent its complex structure."]},{"key":"dc:format","label":"Dc Format","values":["application/pdf"]},{"key":"dc:title","label":"Title","values":["Dirbtinių neuroninių tinklų apmokymas su triukšmu /","Artificial neural network training with noise."]}]}],"canonical_facts":{"dc:creator":["Rinkevičiūtė, Gabrielė,"],"dc:date":["2026"],"dc:description":["This thesis examines the impact of noise in training data on the performance of an artificial neural network and the $k$-nearest neighbors method in regression tasks. For the practical experiments, artificial datasets were generated using the Sphere, Rosenbrock, and Rastrigin benchmark functions. Gaussian noise was applied to introduce noise into the data. Practical experiments were conducted using both methods on clean and noisy datasets of different sizes. The performance of the methods was evaluated using the mean absolute error and standard deviation. The results showed that the artificial neural network utilized increasing amounts of data more effectively when approximating the Sphere and Rosenbrock functions. In contrast, performance improvement on the Rastrigin function was limited by the selected architecture with a sigmoid activation function in the hidden layer. Regarding the $k$-nearest neighbors method, a consistent improvement in performance was observed as the dataset size increased. This characteristic allowed the method to achieve better results for the Rastrigin function, which was more difficult for the artificial neural network to approximate. The study also found that the effect of added noise was greater for functions whose structure the artificial neural network learned more accurately. The lower impact of noise in the case of the Rastrigin function was associated with the model’s limited ability to represent its complex structure."],"dc:format":["application/pdf"],"dc:identifier":["https://repository.vu.lt/VU:ELABAETD308337942&prefLang=en_US"],"dc:language":["lit"],"dc:publisher":["Institutional Repository of Vilnius University"],"dc:relation":["https://epublications.vu.lt/object/elaba:308337942/308337942.pdf"],"dc:rights":["info:eu-repo/semantics/openAccess"],"dc:title":["Dirbtinių neuroninių tinklų apmokymas su triukšmu /","Artificial neural network training with noise."],"dc:type":["info:eu-repo/semantics/bachelorThesis"]},"updated_at":"2026-11-19T10:12:59Z"}