{"id":{"repo_id":"ets-quebec","oai_identifier":"oai:espace.etsmtl.ca:3037"},"canonical_url":"https://search.ndltd.org/etd/ets-quebec/oai:espace.etsmtl.ca:3037","repository":{"repo_id":"ets-quebec","name":"ETS (Quebec)","base_url":"https://espace.etsmtl.ca/cgi/oai2"},"display":{"title":"Summarizing code entities discussed in stack overflow using unsupervised learning","abstract":"Automated source code summarization is an approach that generates natural language descriptions that describe the code entities such as methods and classes to better understand the code and to optimize the time and effort of software developers. There are two ways to summarize code: extractive way as well as abstractive way. Multiple methods and algorithms including Information Retrieval techniques and supervised learning have been applied to generate automatic summaries for code entities. Their common shortcoming is that very little to none of those works has offered a completely automated way, without human intervention, to provide accurate summaries for code entities discussed in informal documentation such as emails exchanged between developers, Stack Overflow, etc. In addition, very little research has investigated informal documentation when summarizing code entities discussed in code summarization. In this thesis, I leverage unsupervised learning, i.e., TextRank to extract relevant sentences that should be part of the summary produced for Android methods mentioned in Stack Overflow. In addition, I use Natural Language Processing (NLP) techniques, in particular, Term Frequency-Inverse Document Frequency (TF-IDF) and Stack Overflow’s word embedding, which are used as text vectorization techniques, as two separate techniques for text vectorization, and I show how different vectorization affect the quality of the produced summaries. To evaluated my research work, I conducted an empirical study that involves 3,084,143 Stack Overflow posts from Stack Overflow tagged between 2009 and April 2020 and 20 Android methods. The results of our investigation show that we can produce automatic summaries for code entities discussed in informal documentation using TextRank and that applying a word embedding vectorization provides more stable results than the TF-IDF when applied with TextRank algorithms.","abstract_html":"Automated source code summarization is an approach that generates natural language descriptions that describe the code entities such as methods and classes to better understand the code and to optimize the time and effort of software developers. There are two ways to summarize code: extractive way as well as abstractive way. Multiple methods and algorithms including Information Retrieval techniques and supervised learning have been applied to generate automatic summaries for code entities. Their common shortcoming is that very little to none of those works has offered a completely automated way, without human intervention, to provide accurate summaries for code entities discussed in informal documentation such as emails exchanged between developers, Stack Overflow, etc. In addition, very little research has investigated informal documentation when summarizing code entities discussed in code summarization. In this thesis, I leverage unsupervised learning, i.e., TextRank to extract relevant sentences that should be part of the summary produced for Android methods mentioned in Stack Overflow. In addition, I use Natural Language Processing (NLP) techniques, in particular, Term Frequency-Inverse Document Frequency (TF-IDF) and Stack Overflow’s word embedding, which are used as text vectorization techniques, as two separate techniques for text vectorization, and I show how different vectorization affect the quality of the produced summaries. To evaluated my research work, I conducted an empirical study that involves 3,084,143 Stack Overflow posts from Stack Overflow tagged between 2009 and April 2020 and 20 Android methods. The results of our investigation show that we can produce automatic summaries for code entities discussed in informal documentation using TextRank and that applying a word embedding vectorization provides more stable results than the TF-IDF when applied with TextRank algorithms.","abstract_has_math":false,"creators":["Harirpoosh, Andisheh"],"institution":"École de technologie supérieure","degree_name":"Mémoire de maîtrise électronique","degree_level":"masters","degree_discipline":null,"degree_department":null,"school":null,"contributors":[],"advisors":[],"committee_chairs":[],"committee_members":[],"year":2022,"date_issued":"2022-02-22","date_published":"2022-02-22","updated_at":"2026-11-18T22:09:26Z","subjects":["synthèse du code, entités de code, stackoverflow, apprentissage non supervisé, traitement du langage naturel, documentation informelle"],"languages":["en"],"rights":[],"rights_urls":[],"identifier_entries":[{"key":"dc:identifier","label":"Identifier","values":["INSERT-YOUR-LIBRARY-ID-HERE3037"],"render_values":[{"text":"INSERT-YOUR-LIBRARY-ID-HERE3037","href":null,"code":true}]}],"person_identifier_entries":[]},"links":{"outbound_url":null,"outbound_label":null,"outbound_source":null,"record_url":null,"record_label":null,"record_kind":null,"record_source":null,"download_url":"https://espace.etsmtl.ca/id/eprint/3037/1/HARIRPOOSH_Andisheh_v2.pdf","download_label":"Download PDF","download_source":"dc:identifier","access_status":null,"embargo_until":null,"access_note":null},"source_record":{"url":"https://espace.etsmtl.ca/cgi/oai2?verb=GetRecord&metadataPrefix=oai_etdms&identifier=oai%3Aespace.etsmtl.ca%3A3037","prefix":"oai_etdms"},"metadata_groups":[{"id":"people","label":"People","entries":[{"key":"dc:creator","label":"Author","values":["Harirpoosh, Andisheh"]}]},{"id":"academic_context","label":"Academic Context","entries":[{"key":"dc:date","label":"Dc Date","values":["2022-02-22"]},{"key":"dc:publisher","label":"Institution","values":["École de technologie supérieure"]},{"key":"dc:type","label":"Dc Type","values":["Electronic Thesis or Dissertation"]},{"key":"thesis:degree_level","label":"Degree Level","values":["masters"]},{"key":"thesis:degree_name","label":"Degree Name","values":["Mémoire de maîtrise électronique"]},{"key":"thesis:institution_name","label":"Thesis Institution Name","values":["École de technologie supérieure"]}]},{"id":"subjects_keywords","label":"Subjects and Keywords","entries":[{"key":"dc:subject","label":"Dc Subject","values":["synthèse du code, entités de code, stackoverflow, apprentissage non supervisé, traitement du langage naturel, documentation informelle"]}]},{"id":"language_rights","label":"Language and Rights","entries":[{"key":"dc:language","label":"Dc Language","values":["en"]}]},{"id":"identifiers","label":"Identifiers","entries":[{"key":"dc:identifier","label":"Identifier","values":["INSERT-YOUR-LIBRARY-ID-HERE3037","https://espace.etsmtl.ca/id/eprint/3037/1/HARIRPOOSH_Andisheh_v2.pdf"]}]},{"id":"additional","label":"Additional Metadata","entries":[{"key":"dc:description","label":"Description","values":["Automated source code summarization is an approach that generates natural language descriptions that describe the code entities such as methods and classes to better understand the code and to optimize the time and effort of software developers. There are two ways to summarize code: extractive way as well as abstractive way. Multiple methods and algorithms including Information Retrieval techniques and supervised learning have been applied to generate automatic summaries for code entities. Their common shortcoming is that very little to none of those works has offered a completely automated way, without human intervention, to provide accurate summaries for code entities discussed in informal documentation such as emails exchanged between developers, Stack Overflow, etc. In addition, very little research has investigated informal documentation when summarizing code entities discussed in code summarization. In this thesis, I leverage unsupervised learning, i.e., TextRank to extract relevant sentences that should be part of the summary produced for Android methods mentioned in Stack Overflow. In addition, I use Natural Language Processing (NLP) techniques, in particular, Term Frequency-Inverse Document Frequency (TF-IDF) and Stack Overflow’s word embedding, which are used as text vectorization techniques, as two separate techniques for text vectorization, and I show how different vectorization affect the quality of the produced summaries. To evaluated my research work, I conducted an empirical study that involves 3,084,143 Stack Overflow posts from Stack Overflow tagged between 2009 and April 2020 and 20 Android methods. The results of our investigation show that we can produce automatic summaries for code entities discussed in informal documentation using TextRank and that applying a word embedding vectorization provides more stable results than the TF-IDF when applied with TextRank algorithms."]},{"key":"dc:format","label":"Dc Format","values":["application/pdf"]},{"key":"dc:title","label":"Title","values":["Summarizing code entities discussed in stack overflow using unsupervised learning"]}]}],"canonical_facts":{"dc:creator":["Harirpoosh, Andisheh"],"dc:date":["2022-02-22"],"dc:description":["Automated source code summarization is an approach that generates natural language descriptions that describe the code entities such as methods and classes to better understand the code and to optimize the time and effort of software developers. There are two ways to summarize code: extractive way as well as abstractive way. Multiple methods and algorithms including Information Retrieval techniques and supervised learning have been applied to generate automatic summaries for code entities. Their common shortcoming is that very little to none of those works has offered a completely automated way, without human intervention, to provide accurate summaries for code entities discussed in informal documentation such as emails exchanged between developers, Stack Overflow, etc. In addition, very little research has investigated informal documentation when summarizing code entities discussed in code summarization. In this thesis, I leverage unsupervised learning, i.e., TextRank to extract relevant sentences that should be part of the summary produced for Android methods mentioned in Stack Overflow. In addition, I use Natural Language Processing (NLP) techniques, in particular, Term Frequency-Inverse Document Frequency (TF-IDF) and Stack Overflow’s word embedding, which are used as text vectorization techniques, as two separate techniques for text vectorization, and I show how different vectorization affect the quality of the produced summaries. To evaluated my research work, I conducted an empirical study that involves 3,084,143 Stack Overflow posts from Stack Overflow tagged between 2009 and April 2020 and 20 Android methods. The results of our investigation show that we can produce automatic summaries for code entities discussed in informal documentation using TextRank and that applying a word embedding vectorization provides more stable results than the TF-IDF when applied with TextRank algorithms."],"dc:format":["application/pdf"],"dc:identifier":["INSERT-YOUR-LIBRARY-ID-HERE3037","https://espace.etsmtl.ca/id/eprint/3037/1/HARIRPOOSH_Andisheh_v2.pdf"],"dc:language":["en"],"dc:publisher":["École de technologie supérieure"],"dc:subject":["synthèse du code, entités de code, stackoverflow, apprentissage non supervisé, traitement du langage naturel, documentation informelle"],"dc:title":["Summarizing code entities discussed in stack overflow using unsupervised learning"],"dc:type":["Electronic Thesis or Dissertation"],"thesis:degree_level":["masters"],"thesis:degree_name":["Mémoire de maîtrise électronique"],"thesis:institution_name":["École de technologie supérieure"]},"updated_at":"2026-11-18T22:09:26Z"}