[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"doc-detail-119186-en":3,"doc-seo-119186-105":30,"detail-sidebar-cat-0-en-105":96},{"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":20,"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},119186,8796095461610,"Oliver","https://ap-avatar.wpscdn.com/davatar_276721f389ce27ea32af1340a28f341c",8,"Research & Report","Deep Behavioral Analysis of Machine Learning Algorithms Against Data Poisoning - Research findings","Poisoning attacks pose a practical threat to machine learning systems by infiltrating training data and degrading model integrity. This work provides a deep behavioral analysis of six ML algorithms, measuring how poisoning levels correlate with classification accuracy. Using public datasets (UNSW-NB15, BotDroid, CTU13, CIC-IDS-2017) and poisoning rates from 5% to 25%, the study evaluates accuracy, precision, recall, F1-score, false positive rate, and ROC, supported by sensitivity experiments across algorithms and feature-noise effects.","Deep Behavioral Analysis of Machine Learning Algorithms Against Data Poisoning  \nAnum Paracha, Junaid Arshad, Mohamed Ben Farah, Khalid Ismail  \na Department of Computing, Birmingham City University, UK  \nAbstract  \nPoisoning attacks represent one of the most common and practical adversarial attempts on machine learning systems. In this paper, we have conducted a deep behavioural analysis of six machine learning (ML) algorithms, analyzing poisoning impact and correlation between poisoning levels and classification accuracy. Adopting an empirical approach, we highlight practical feasibility of data poisoning, comprehensively analyzing factors of individual algorithms affected by poisoning. We used public datasets (UNSW-NB15, BotDroid, CTU13, and CIC-IDS-2017) and varying poisoning levels (5% - 25%) to conduct rigorous analysis across different settings. In particular, we analyzed the accuracy, precision, recall, f1-score, false positive rate and ROC of the chosen algorithms. Further, we conducted a sensitivity analysis of each algorithm to understand the impact of poisoning on its performance and characteristics underpinning its susceptibility against data poisoning attacks. Our analysis shows that, for 15% poisoning of UNSW-NB15 dataset, the accuracy of Decision Tree (DT) decreases by 15.04% with an increase of 14.85% in false positive rate. Further, with 25% poisoning of BotDroid dataset, accuracy of K-nearest neighbours (KNN) decreases by 15.48% . On the other hand, Random Forest (RF) is comparatively more resilient against poisoned training data with a decrease of 8.5% inaccuracy with 15% poisoning of UNSW-NB15 dataset and 5.2% for BotDroid dataset. Our results highlight that 10%-15% of dataset poisoning is the most effective poisoning rate, significantly disrupting classifiers without introducing overfitting, whereas 25% is detectable because of high performance degradation and overfitting algorithms. Our analysis also helps understand how asymmetric features and noise affect the impact of data poisoning on machine learning classifiers. Our experimentation and analysis are publicly available at: [https:](https://github.com/AnumAtique/Behavioural-Analaysis-of-Poisoned-ML/)[//](https://github.com/AnumAtique/Behavioural-Analaysis-of-Poisoned-ML/)[github.com](https://github.com/AnumAtique/Behavioural-Analaysis-of-Poisoned-ML/)[/](https://github.com/AnumAtique/Behavioural-Analaysis-of-Poisoned-ML/)[AnumAtique](https://github.com/AnumAtique/Behavioural-Analaysis-of-Poisoned-ML/)[/](https://github.com/AnumAtique/Behavioural-Analaysis-of-Poisoned-ML/)[Behavioural-Analaysis-of-Poisoned-ML](https://github.com/AnumAtique/Behavioural-Analaysis-of-Poisoned-ML/)[/](https://github.com/AnumAtique/Behavioural-Analaysis-of-Poisoned-ML/)  \nKeywords: Behavioral Analysis, Adversarial Poisoning, Integrity Violation, Complacent Poisoning  \n1. Introduction  \nMachine learning models are widespread, facilitating cuttingedge digital solutions in a range of scenarios including securitycritical applications such as malware detector[1], intrusion detection system[5], automated firewalls[6], and biometric recognition[7] . Machine learning models help understand patterns from the given dataset and train themselves to predict and classify new data without requiring additional information or interaction with any third party such as humans. With the training dataset, machine learning models develop the dynamic classification mechanism that leverages these models to understand the nature of new data and classify them.  \nSuch proliferation of machine learning models and their dynamic classification mechanism renders their security fundamental to the security of systems underpinned by them. Several attacks have been explored in literature aiming to compromise the performance and accuracy of machine learning algorithms such as [8], [9], [10] and [11] . Among these attacks, data poisoning [8] is one of the most prominent attacks on machine learning whereby an ad","cbCaieX9Wo7dENEN","https://ap.wps.com/l/cbCaieX9Wo7dENEN","pdf",1925334,1,18,"English","en",105,"# Abstract\n# Introduction\n## Background and relevance of data poisoning\n## Poisoning techniques and real-world impact\n## Motivation for behavioral analysis","[{\"question\":\"What does the paper analyze about data poisoning attacks?\",\"answer\":\"It analyzes the behavioral impact of data poisoning on six machine learning algorithms, focusing on how poisoning levels affect classification accuracy and related performance characteristics.\"},{\"question\":\"Which datasets and poisoning rates are used in the experiments?\",\"answer\":\"The experiments use UNSW-NB15, BotDroid, CTU13, and CIC-IDS-2017, with poisoning levels varying from 5% to 25%.\"},{\"question\":\"How do the results compare the resilience of different algorithms?\",\"answer\":\"Random Forest shows comparatively better resilience, while Decision Tree and KNN experience larger accuracy drops under specific poisoning settings (e.g., DT on UNSW-NB15 at 15%).\"},{\"question\":\"What poisoning rate range is highlighted as most effective and why?\",\"answer\":\"The study indicates that 10%–15% poisoning most effectively disrupts classifiers without introducing overfitting, while 25% is detectable due to substantial performance degradation and overfitting.\"}]","Deep Behavioral Analysis of Machine Learning Algorithms Against Data Poisoning - Research findings | PDF",1785722985,45,{"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":91,"head_meta":93,"extra_data":95,"updated_unix":28},"deep-behavioral-analysis-of-machine-learning-algorithms-against-data-poisoning-research-findings","",{"@graph":36,"@context":90},[37,54,69],{"@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/deep-behavioral-analysis-of-machine-learning-algorithms-against-data-poisoning-research-findings/119186/",4,{"url":52,"name":13,"@type":55,"author":56,"headline":13,"publisher":58,"fileFormat":61,"inLanguage":23,"description":14,"dateModified":62,"datePublished":63,"encodingFormat":61,"isAccessibleForFree":64,"interactionStatistic":65},"DigitalDocument",{"name":9,"@type":57},"Person",{"url":41,"name":59,"@type":60},"DocShare","Organization","application/pdf","2026-08-04","2026-08-03",true,{"@type":66,"interactionType":67,"userInteractionCount":20},"InteractionCounter",{"@type":68},"ViewAction",{"@type":70,"mainEntity":71},"FAQPage",[72,78,82,86],{"name":73,"@type":74,"acceptedAnswer":75},"What does the paper analyze about data poisoning attacks?","Question",{"text":76,"@type":77},"It analyzes the behavioral impact of data poisoning on six machine learning algorithms, focusing on how poisoning levels affect classification accuracy and related performance characteristics.","Answer",{"name":79,"@type":74,"acceptedAnswer":80},"Which datasets and poisoning rates are used in the experiments?",{"text":81,"@type":77},"The experiments use UNSW-NB15, BotDroid, CTU13, and CIC-IDS-2017, with poisoning levels varying from 5% to 25%.",{"name":83,"@type":74,"acceptedAnswer":84},"How do the results compare the resilience of different algorithms?",{"text":85,"@type":77},"Random Forest shows comparatively better resilience, while Decision Tree and KNN experience larger accuracy drops under specific poisoning settings (e.g., DT on UNSW-NB15 at 15%).",{"name":87,"@type":74,"acceptedAnswer":88},"What poisoning rate range is highlighted as most effective and why?",{"text":89,"@type":77},"The study indicates that 10%–15% poisoning most effectively disrupts classifiers without introducing overfitting, while 25% is detectable due to substantial performance degradation and overfitting.","https://schema.org",{"og:url":52,"og:type":92,"og:title":13,"og:site_name":59,"og:description":14},"article",{"robots":94,"canonical":52},"index,follow",{"doc_id":7,"site_id":24},{"code":4,"msg":5,"data":97},[98,102,106,110,115,120,125,128,133,136,140],{"id":20,"doc_module":4,"doc_module_name":46,"category_name":99,"show_sort_weight":100,"slug":101},"Story & Novel",90,"story-novel",{"id":47,"doc_module":4,"doc_module_name":46,"category_name":103,"show_sort_weight":104,"slug":105},"Literature",80,"literature",{"id":53,"doc_module":4,"doc_module_name":46,"category_name":107,"show_sort_weight":108,"slug":109},"Exam",70,"exam",{"id":111,"doc_module":4,"doc_module_name":46,"category_name":112,"show_sort_weight":113,"slug":114},5,"Comic",60,"comic",{"id":116,"doc_module":4,"doc_module_name":46,"category_name":117,"show_sort_weight":118,"slug":119},6,"Technology",50,"technology",{"id":121,"doc_module":4,"doc_module_name":46,"category_name":122,"show_sort_weight":123,"slug":124},7,"Healthcare",40,"healthcare",{"id":11,"doc_module":4,"doc_module_name":46,"category_name":12,"show_sort_weight":126,"slug":127},30,"research-report",{"id":129,"doc_module":4,"doc_module_name":46,"category_name":130,"show_sort_weight":131,"slug":132},9,"Religion & Spirituality",20,"religion-spirituality",{"id":131,"doc_module":4,"doc_module_name":46,"category_name":134,"show_sort_weight":131,"slug":135},"World Cup","world-cup",{"id":137,"doc_module":4,"doc_module_name":46,"category_name":138,"show_sort_weight":137,"slug":139},10,"Lifestyle","lifestyle",{"id":141,"doc_module":4,"doc_module_name":46,"category_name":142,"show_sort_weight":111,"slug":143},19,"General","general"]