[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"doc-detail-118943-en":3,"doc-seo-118943-105":30,"detail-sidebar-cat-0-en-105":91},{"code":4,"msg":5,"data":6},0,"success",{"doc_id":7,"user_id":8,"nickname":9,"user_avatar":10,"doc_module":4,"category_id":11,"category_name":12,"doc_title":13,"doc_description":14,"doc_content":15,"file_id":16,"file_url":17,"file_type":18,"file_size":19,"view_count":4,"is_deleted":4,"is_public":20,"is_downloadable":20,"audit_status":20,"page_count":21,"language":22,"language_code":23,"site_id":24,"html_lang":23,"table_of_contents":25,"faqs":26,"seo_title":27,"seo_description":14,"update_tm":28,"read_time":29},118943,3848291630094,"Emma Wilson","https://eur-avatar.wpscdn.com/davatar_085a072bc5b1113ac321206ff7593b45",8,"Research & Report","Delete My Account - Impact of Data Deletion on Machine Learning Classifiers","User data is increasingly protected as security breaches and privacy incidents have become more visible, and GDPR’s right to erasure has pushed many individuals to act on deletion requests. This work analyzes how exercising the right to erasure affects machine learning performance on classification tasks. Using multiple datasets and learning algorithms, experiments cover different deletion-behavior scenarios, with reasonable assumptions to address limited data on real user behavior. Results show strong dependence on deletion amount, dataset characteristics, and chosen bias assumptions regarding user behavior and data quality.","Delete My Account: Impact of Data Deletion on Machine Learning Classifiers  \nTobias Dam  \nInstitute of IT Security Research St. Pölten UAS  \nSt. Pölten, Austria  \n[tobias.dam@fhstp.ac.at](tobias.dam@fhstp.ac.at)  \nMaximilian Henzl  \nInstitute of IT Security Research St. Pölten UAS  \nSt. Pölten, Austria  \n[is201849@fhstp.ac.at](is201849@fhstp.ac.at)  \nLukas Daniel Klausner  \nInstitute of IT Security Research St. Pölten UAS  \nSt. Pölten, Austria  \n[mail@l17r.eu](mail@l17r.eu)  \narXiv :2311 . 10385v1 [ cs .LG] 17 Nov 2023  \nAbstract—Users are more aware than ever of the importance of their own data, thanks to reports about security breaches and leaks of private, often sensitive data in recent years. Additionally, the GDPR has been in effect in the European Union for over three years and many people have encountered its effects in oneway or another. Consequently, more and more users are actively protecting their personal data. One way to do this is to make of the right to erasure guaranteed in the GDPR, which has potential implications for a number of different fields, such as big data and machine learning.  \nOur paper presents an in-depth analysis about the impact of the use of the right to erasure on the performance of machine learning models on classification tasks. We conduct various experiments utilising different datasets as well as different machine learning algorithms to analyse a variety of deletion behaviour scenarios. Due to the lack of credible data on actual user behaviour, we make reasonable assumptions for various deletion modes and biases and provide insight into the effects of different plausible scenarios for right to erasure usage on data quality of machine learning. Our results show that the impact depends strongly on the amount of data deleted, the particular characteristics of the dataset and the bias chosen for deletion and assumptions on user behaviour.  \nIndex Terms—deletion request; machine learning; right to erasure  \nI. INTRODUCTION  \nThe General Data Protection Regulation (GDPR) [1], in effect since 25 May 2018, was a significant development for improving privacy and protecting the data of EU citizens. Among various rights granted to citizens regarding their data (such asthe right of access by the data subject), the GDPR also grants the right to erasure. The right to erasure, also called the right to be forgotten, empowers citizens to request the deletion of their personal data, with only a limited set of explicitly defined exceptions.  \nPersonal data as defined in the GDPR (as “information relating to a natural person”) includes identifiers, such as names, but also characteristics that may indirectly identify a person, such as data on their economical or social identity. In recent years, the amount of personal information that is collected, stored and processed has been steadily increasing. More and more companies apply concepts like big data [2], [3] as well  \nas machine learning in order to automatically process large amounts of collected data in order to provide personalised services, sell targeted advertising or conduct research.  \nAs privacy awareness is continuously increasing over the past few years and the right to erasure becomes more wellknown, people are more likely to make use of their rights and request the deletion of their data. By way of example, the announcement of a modification to the privacy policy of WhatsApp in January 2021 [4] caused millions of users to switch to other messenger services. It stands to reason that at least part of these users also deleted their accounts and made use of their right to erasure.  \nSuch increases of the usage rates of the right to erasure and the resulting loss of data in datasets used for machine learning tasks might impact the quality of those results. So far, little research has been conducted on the impact of the use of the right to erasure on machine learning. The primary aim of this paper is to investigate the influence of different amounts an","cbCaiiswjA8eJtcq","https://ap.wps.com/l/cbCaiiswjA8eJtcq","pdf",762807,1,14,"English","en",105,"# Introduction\n## GDPR and the right to erasure\n## Data protection awareness and account deletion\n## Goal and contributions\n# Related Work\n# Methodology\n## Deletion scenarios and assumptions","[{\"question\":\"What is the right to erasure and why is it relevant to machine learning?\",\"answer\":\"The right to erasure allows individuals to request deletion of personal data under GDPR, with limited exceptions. Removing data can change dataset composition and therefore affect how machine learning models perform on classification tasks.\"},{\"question\":\"How does the paper study the impact of data deletion?\",\"answer\":\"It runs experiments using different datasets and machine learning algorithms, defining multiple deletion-behavior scenarios. The study varies deletion likelihood scenarios and the percentage of deleted records to analyze performance changes.\"},{\"question\":\"What factors most strongly influence the impact on classifier performance?\",\"answer\":\"The results indicate the impact depends heavily on how much data is deleted, characteristics of the dataset, and the bias and assumptions used for deletion behavior.\"}]","Delete My Account - Impact of Data Deletion on Machine Learning Classifiers | PDF",1785721125,35,{"code":4,"msg":31,"data":32},"ok",{"site_id":24,"language":23,"slug":33,"title":13,"keywords":34,"description":14,"schema_data":35,"social_meta":86,"head_meta":88,"extra_data":90,"updated_unix":28},"delete-my-account-impact-of-data-deletion-on-machine-learning-classifiers","",{"@graph":36,"@context":85},[37,54,68],{"@type":38,"itemListElement":39},"BreadcrumbList",[40,44,48,51],{"item":41,"name":42,"@type":43,"position":20},"https://docshare.wps.com","Home","ListItem",{"item":45,"name":46,"@type":43,"position":47},"https://docshare.wps.com/document/","Document",2,{"item":49,"name":12,"@type":43,"position":50},"https://docshare.wps.com/document/research-report/",3,{"item":52,"name":13,"@type":43,"position":53},"https://docshare.wps.com/document/delete-my-account-impact-of-data-deletion-on-machine-learning-classifiers/118943/",4,{"url":52,"name":13,"@type":55,"author":56,"headline":13,"publisher":58,"fileFormat":61,"inLanguage":23,"description":14,"dateModified":62,"datePublished":62,"encodingFormat":61,"isAccessibleForFree":63,"interactionStatistic":64},"DigitalDocument",{"name":9,"@type":57},"Person",{"url":41,"name":59,"@type":60},"DocShare","Organization","application/pdf","2026-08-03",true,{"@type":65,"interactionType":66,"userInteractionCount":4},"InteractionCounter",{"@type":67},"ViewAction",{"@type":69,"mainEntity":70},"FAQPage",[71,77,81],{"name":72,"@type":73,"acceptedAnswer":74},"What is the right to erasure and why is it relevant to machine learning?","Question",{"text":75,"@type":76},"The right to erasure allows individuals to request deletion of personal data under GDPR, with limited exceptions. Removing data can change dataset composition and therefore affect how machine learning models perform on classification tasks.","Answer",{"name":78,"@type":73,"acceptedAnswer":79},"How does the paper study the impact of data deletion?",{"text":80,"@type":76},"It runs experiments using different datasets and machine learning algorithms, defining multiple deletion-behavior scenarios. The study varies deletion likelihood scenarios and the percentage of deleted records to analyze performance changes.",{"name":82,"@type":73,"acceptedAnswer":83},"What factors most strongly influence the impact on classifier performance?",{"text":84,"@type":76},"The results indicate the impact depends heavily on how much data is deleted, characteristics of the dataset, and the bias and assumptions used for deletion behavior.","https://schema.org",{"og:url":52,"og:type":87,"og:title":13,"og:site_name":59,"og:description":14},"article",{"robots":89,"canonical":52},"index,follow",{"doc_id":7,"site_id":24},{"code":4,"msg":5,"data":92},[93,97,101,105,110,115,120,123,128,131,135],{"id":20,"doc_module":4,"doc_module_name":46,"category_name":94,"show_sort_weight":95,"slug":96},"Story & Novel",90,"story-novel",{"id":47,"doc_module":4,"doc_module_name":46,"category_name":98,"show_sort_weight":99,"slug":100},"Literature",80,"literature",{"id":53,"doc_module":4,"doc_module_name":46,"category_name":102,"show_sort_weight":103,"slug":104},"Exam",70,"exam",{"id":106,"doc_module":4,"doc_module_name":46,"category_name":107,"show_sort_weight":108,"slug":109},5,"Comic",60,"comic",{"id":111,"doc_module":4,"doc_module_name":46,"category_name":112,"show_sort_weight":113,"slug":114},6,"Technology",50,"technology",{"id":116,"doc_module":4,"doc_module_name":46,"category_name":117,"show_sort_weight":118,"slug":119},7,"Healthcare",40,"healthcare",{"id":11,"doc_module":4,"doc_module_name":46,"category_name":12,"show_sort_weight":121,"slug":122},30,"research-report",{"id":124,"doc_module":4,"doc_module_name":46,"category_name":125,"show_sort_weight":126,"slug":127},9,"Religion & Spirituality",20,"religion-spirituality",{"id":126,"doc_module":4,"doc_module_name":46,"category_name":129,"show_sort_weight":126,"slug":130},"World Cup","world-cup",{"id":132,"doc_module":4,"doc_module_name":46,"category_name":133,"show_sort_weight":132,"slug":134},10,"Lifestyle","lifestyle",{"id":136,"doc_module":4,"doc_module_name":46,"category_name":137,"show_sort_weight":106,"slug":138},19,"General","general"]