{"id":{"repo_id":"vilnius","oai_identifier":"oai:vu.lt:elaba:308336676"},"canonical_url":"https://search.ndltd.org/etd/vilnius/oai:vu.lt:elaba:308336676","repository":{"repo_id":"vilnius","name":"Vilnius University","base_url":"https://epublications.vu.lt/oai"},"display":{"title":"Design and implementation of a tool aimed at identifying defect-prone functions using combined static, dynamic and evolutionary analysis /","abstract":"Software defect prediction helps developers focus verification effort on the parts of a system most likely to contain faults. Existing tools typically rely on static and evolutionary metrics, rarely combine them with dynamic information such as test coverage and most often operate at the file or class level, which is too coarse to direct a developer to a specific function. Python, despite its widespread use, has received comparatively little attention in this domain. This thesis addresses these gaps by designing and implementing a function-level defect prediction tool for Python. The tool integrates five metrics: lines of code, cyclomatic complexity, coupling, code churn and test coverage, into a ranked risk score. Three models were trained and evaluated on a labeled dataset of 6195 functions from 16 open-source Python projects in the BugsInPy benchmark: a weighted linear scoring model trained with Differential Evolution, Logistic Regression and Random Forest. Under Leave-One-Project-Out cross-validation, the weighted linear scoring model trained with Differential Evolution achieved the best generalisation with AUC of 0.699 and Lift of 1.96 and was selected for the final tool. The learned weights identify code churn as the dominant predictor with 0.571, followed by lines of code with 0.183 and cyclomatic complexity with 0.167. The resulting command-line tool accepts a repository path and an optional coverage report, and outputs a ranked list of defect-prone functions configurable through command-line flags.","abstract_html":"Software defect prediction helps developers focus verification effort on the parts of a system most likely to contain faults. Existing tools typically rely on static and evolutionary metrics, rarely combine them with dynamic information such as test coverage and most often operate at the file or class level, which is too coarse to direct a developer to a specific function. Python, despite its widespread use, has received comparatively little attention in this domain. This thesis addresses these gaps by designing and implementing a function-level defect prediction tool for Python. The tool integrates five metrics: lines of code, cyclomatic complexity, coupling, code churn and test coverage, into a ranked risk score. Three models were trained and evaluated on a labeled dataset of 6195 functions from 16 open-source Python projects in the BugsInPy benchmark: a weighted linear scoring model trained with Differential Evolution, Logistic Regression and Random Forest. Under Leave-One-Project-Out cross-validation, the weighted linear scoring model trained with Differential Evolution achieved the best generalisation with AUC of 0.699 and Lift of 1.96 and was selected for the final tool. The learned weights identify code churn as the dominant predictor with 0.571, followed by lines of code with 0.183 and cyclomatic complexity with 0.167. The resulting command-line tool accepts a repository path and an optional coverage report, and outputs a ranked list of defect-prone functions configurable through command-line flags.","abstract_has_math":false,"creators":["Drungilaitė, 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":["software defect prediction, code churn, static code metrics, function level analysis, linear model"],"languages":["eng"],"rights":["info:eu-repo/semantics/openAccess"],"rights_urls":[],"identifier_entries":[],"person_identifier_entries":[]},"links":{"outbound_url":"https://repository.vu.lt/VU:ELABAETD308336676&prefLang=en_US","outbound_label":"Repository record","outbound_source":"dc:identifier","record_url":"https://repository.vu.lt/VU:ELABAETD308336676&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%3A308336676","prefix":"oai_dc"},"metadata_groups":[{"id":"people","label":"People","entries":[{"key":"dc:creator","label":"Author","values":["Drungilaitė, 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:308336676/308336676.pdf"]},{"key":"dc:type","label":"Dc Type","values":["info:eu-repo/semantics/bachelorThesis"]}]},{"id":"subjects_keywords","label":"Subjects and Keywords","entries":[{"key":"dc:subject","label":"Dc Subject","values":["software defect prediction, code churn, static code metrics, function level analysis, linear model"]}]},{"id":"language_rights","label":"Language and Rights","entries":[{"key":"dc:language","label":"Dc Language","values":["eng"]},{"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:ELABAETD308336676&prefLang=en_US"]}]},{"id":"additional","label":"Additional Metadata","entries":[{"key":"dc:description","label":"Description","values":["Software defect prediction helps developers focus verification effort on the parts of a system most likely to contain faults. Existing tools typically rely on static and evolutionary metrics, rarely combine them with dynamic information such as test coverage and most often operate at the file or class level, which is too coarse to direct a developer to a specific function. Python, despite its widespread use, has received comparatively little attention in this domain. This thesis addresses these gaps by designing and implementing a function-level defect prediction tool for Python. The tool integrates five metrics: lines of code, cyclomatic complexity, coupling, code churn and test coverage, into a ranked risk score. Three models were trained and evaluated on a labeled dataset of 6195 functions from 16 open-source Python projects in the BugsInPy benchmark: a weighted linear scoring model trained with Differential Evolution, Logistic Regression and Random Forest. Under Leave-One-Project-Out cross-validation, the weighted linear scoring model trained with Differential Evolution achieved the best generalisation with AUC of 0.699 and Lift of 1.96 and was selected for the final tool. The learned weights identify code churn as the dominant predictor with 0.571, followed by lines of code with 0.183 and cyclomatic complexity with 0.167. The resulting command-line tool accepts a repository path and an optional coverage report, and outputs a ranked list of defect-prone functions configurable through command-line flags."]},{"key":"dc:format","label":"Dc Format","values":["application/pdf"]},{"key":"dc:title","label":"Title","values":["Design and implementation of a tool aimed at identifying defect-prone functions using combined static, dynamic and evolutionary analysis /","Įrankio, skirto atrasti potencialiai defektus turinčias funkcijas, derinant statinę, dinaminę ir evoliucinę analizę, projektavimas ir įgyvendinimas."]}]}],"canonical_facts":{"dc:creator":["Drungilaitė, Gabrielė,"],"dc:date":["2026"],"dc:description":["Software defect prediction helps developers focus verification effort on the parts of a system most likely to contain faults. Existing tools typically rely on static and evolutionary metrics, rarely combine them with dynamic information such as test coverage and most often operate at the file or class level, which is too coarse to direct a developer to a specific function. Python, despite its widespread use, has received comparatively little attention in this domain. This thesis addresses these gaps by designing and implementing a function-level defect prediction tool for Python. The tool integrates five metrics: lines of code, cyclomatic complexity, coupling, code churn and test coverage, into a ranked risk score. Three models were trained and evaluated on a labeled dataset of 6195 functions from 16 open-source Python projects in the BugsInPy benchmark: a weighted linear scoring model trained with Differential Evolution, Logistic Regression and Random Forest. Under Leave-One-Project-Out cross-validation, the weighted linear scoring model trained with Differential Evolution achieved the best generalisation with AUC of 0.699 and Lift of 1.96 and was selected for the final tool. The learned weights identify code churn as the dominant predictor with 0.571, followed by lines of code with 0.183 and cyclomatic complexity with 0.167. The resulting command-line tool accepts a repository path and an optional coverage report, and outputs a ranked list of defect-prone functions configurable through command-line flags."],"dc:format":["application/pdf"],"dc:identifier":["https://repository.vu.lt/VU:ELABAETD308336676&prefLang=en_US"],"dc:language":["eng"],"dc:publisher":["Institutional Repository of Vilnius University"],"dc:relation":["https://epublications.vu.lt/object/elaba:308336676/308336676.pdf"],"dc:rights":["info:eu-repo/semantics/openAccess"],"dc:subject":["software defect prediction, code churn, static code metrics, function level analysis, linear model"],"dc:title":["Design and implementation of a tool aimed at identifying defect-prone functions using combined static, dynamic and evolutionary analysis /","Įrankio, skirto atrasti potencialiai defektus turinčias funkcijas, derinant statinę, dinaminę ir evoliucinę analizę, projektavimas ir įgyvendinimas."],"dc:type":["info:eu-repo/semantics/bachelorThesis"]},"updated_at":"2026-11-19T10:12:59Z"}