[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"detail-sidebar-cat-0-en-105":3,"doc-seo-291239-105":59,"doc-detail-291239-en":131},{"code":4,"msg":5,"data":6},0,"success",[7,13,18,23,28,33,38,43,48,51,55],{"id":8,"doc_module":4,"doc_module_name":9,"category_name":10,"show_sort_weight":11,"slug":12},1,"Document","Story & Novel",90,"story-novel",{"id":14,"doc_module":4,"doc_module_name":9,"category_name":15,"show_sort_weight":16,"slug":17},2,"Literature",80,"literature",{"id":19,"doc_module":4,"doc_module_name":9,"category_name":20,"show_sort_weight":21,"slug":22},4,"Exam",70,"exam",{"id":24,"doc_module":4,"doc_module_name":9,"category_name":25,"show_sort_weight":26,"slug":27},5,"Comic",60,"comic",{"id":29,"doc_module":4,"doc_module_name":9,"category_name":30,"show_sort_weight":31,"slug":32},6,"Technology",50,"technology",{"id":34,"doc_module":4,"doc_module_name":9,"category_name":35,"show_sort_weight":36,"slug":37},7,"Healthcare",40,"healthcare",{"id":39,"doc_module":4,"doc_module_name":9,"category_name":40,"show_sort_weight":41,"slug":42},8,"Research & Report",30,"research-report",{"id":44,"doc_module":4,"doc_module_name":9,"category_name":45,"show_sort_weight":46,"slug":47},9,"Religion & Spirituality",20,"religion-spirituality",{"id":46,"doc_module":4,"doc_module_name":9,"category_name":49,"show_sort_weight":46,"slug":50},"World Cup","world-cup",{"id":52,"doc_module":4,"doc_module_name":9,"category_name":53,"show_sort_weight":52,"slug":54},10,"Lifestyle","lifestyle",{"id":56,"doc_module":4,"doc_module_name":9,"category_name":57,"show_sort_weight":24,"slug":58},19,"General","general",{"code":4,"msg":60,"data":61},"ok",{"site_id":62,"language":63,"slug":64,"title":65,"keywords":66,"description":67,"schema_data":68,"social_meta":124,"head_meta":126,"extra_data":128,"updated_unix":130},105,"en","hate-speech-detection-from-transliterated-amharic-social-media-comments-using-machine-learning-and-deep-learning-approaches","Hate Speech Detection from Transliterated Amharic Social Media Comments Using Machine Learning and Deep Learning Approaches","","Transliterated Amharic text on social media challenges hate speech detection models trained on formal-script datasets. This research evaluates how auto-transliterated and manually transliterated data, merged with an existing Amharic hate speech dataset, affect machine learning and deep learning classifiers. Using 3,000 samples with an 80:20 train-test split, experiments compare SVM and single- and multichannel CNN models. Results show multichannel CNN achieves the best F1-score on the existing dataset, while mixed-domain transliteration combinations fail to improve performance, indicating inconsistency-driven domain effects.",{"@graph":69,"@context":123},[70,84,106],{"@type":71,"itemListElement":72},"BreadcrumbList",[73,77,79,82],{"item":74,"name":75,"@type":76,"position":8},"https://docshare.wps.com","Home","ListItem",{"item":78,"name":9,"@type":76,"position":14},"https://docshare.wps.com/document/",{"item":80,"name":40,"@type":76,"position":81},"https://docshare.wps.com/document/research-report/",3,{"item":83,"name":65,"@type":76,"position":19},"https://docshare.wps.com/document/hate-speech-detection-from-transliterated-amharic-social-media-comments-using-machine-learning-and-deep-learning-approaches/291239/",{"url":83,"name":65,"@type":85,"image":86,"author":91,"headline":65,"publisher":94,"fileFormat":97,"inLanguage":63,"description":67,"dateModified":98,"datePublished":99,"encodingFormat":97,"isAccessibleForFree":100,"interactionStatistic":101},"DigitalDocument",{"url":87,"@type":88,"width":89,"height":90},"https://docshare.wps.com/thumbnails/hate-speech-detection-from-transliterated-amharic-social-media-comments-using-machine-learning-and-deep-learning-approaches/291239.png","ImageObject",300,407,{"name":92,"@type":93},"Ophelia","Person",{"url":74,"name":95,"@type":96},"DocShare","Organization","application/pdf","2026-10-28","2026-09-17",true,{"@type":102,"interactionType":103,"userInteractionCount":105},"InteractionCounter",{"@type":104},"ViewAction",14,{"@type":107,"mainEntity":108},"FAQPage",[109,115,119],{"name":110,"@type":111,"acceptedAnswer":112},"What research gap does the study address about Amharic hate speech detection?","Question",{"text":113,"@type":114},"The study targets the lack of research on detecting hate speech from transliterated Amharic comments, especially those posted on social media, and evaluates how auto- vs manually transliterated data affects model performance.","Answer",{"name":116,"@type":111,"acceptedAnswer":117},"Which classifiers are evaluated in the study?",{"text":118,"@type":114},"The study evaluates Support Vector Machine (SVM) and convolutional neural network (CNN) approaches, including single-channel and multichannel CNN models.",{"name":120,"@type":111,"acceptedAnswer":121},"What is the key finding about using transliterated datasets together with the existing dataset?",{"text":122,"@type":114},"Combining transliterated datasets with the existing Amharic dataset does not improve classifier performance, likely due to inconsistencies in transliteration and dataset domain dependencies.","https://schema.org",{"og:url":83,"og:type":125,"og:title":65,"og:site_name":95,"og:description":67},"article",{"robots":127,"canonical":83},"index,follow",{"doc_id":129,"site_id":62},291239,1789643844,{"code":4,"msg":5,"data":132},{"doc_id":129,"user_id":133,"nickname":92,"user_avatar":134,"doc_module":4,"category_id":39,"category_name":40,"doc_title":65,"doc_description":67,"doc_content":135,"file_id":136,"file_url":137,"file_type":138,"file_size":139,"view_count":105,"is_deleted":4,"is_public":8,"is_downloadable":8,"audit_status":8,"page_count":140,"language":141,"language_code":63,"site_id":62,"html_lang":63,"table_of_contents":142,"faqs":143,"seo_title":144,"seo_description":67,"update_tm":130,"read_time":145,"preview_page_count":4},7971461741311,"https://ap-avatar.wpscdn.com/avatar/74000253aff267980c6?x-image-process=image/resize,m_fixed,w_180,h_180&k=1779345379180704826","Journal of Computational Science & Data Analytics © AASTU Press  \nJCSDA, Vol. 1, No. 2, 59–75  \nDOI: 10.69660/jcsda.01022404  \nISSN 2959-6912  \nHate Speech Detection from Transliterated Amharic Social Media Comments Using Machine Learning and Deep Learning Approaches  \nZeleke Abebaw  \nDepartment of Software Engineering  \nArtificial Intelligence and Robotics Center of Excellence  \nAddis Ababa Science and Technology University, Addis Ababa, Ethiopia  \nCorresponding author: [zeleke.abebaw@aastu. edu. et](zeleke.abebaw@aastu. edu. et)  \nAndreas Rauber  \nInstitute of Information Systems Engineering,  \nTechnical University of Vienna, Vienna, Austria  \nSolomon Atnafu  \nDepartment of Computer Science, Addis Ababa University,  \nAddis Ababa, Ethiopia  \nThe rise of transliterated script usage on social media has presented significant challenges to hate speech detection models, as such scripts often bypass models trained exclusively on formal language datasets. Existing Amharic hate speech detection studies predominantly focus on datasets written in formal Amharic scripts using machine learning approaches, leaving transliterated comments underexplored. This research addresses the gap by evaluating the impact of auto-transliterated and manually transliterated datasets, merged with an existing Amharic hate speech dataset, on the performance of machine learning and deep learning classifiers. The study employed a total of 3,000 datasets which is split into ratio of 80:20 for training and testing. The dataset consists of auto-transliterated, manually transliterated, formal Amharic script, and their combinations. The classifiers including Support Vector Machine, single and multichannel Convolutional Neural Networks were assessed. Experimental results show that the multichannel CNN outperformed single-channel CNN models on the existing Amharic dataset, achieving an F1-score of 0.810 compared to 0.783 and 0.769 for single channel and multichannel CNN, respectively. However, combining transliterated datasets with the existing dataset did not improve classifier performance, likely due to the inconsistencies in scrip transliteration and dataset domain dependencies. This study concludes that transliterated datasets should be treated separately for hate speech detection, and combining datasets from different domains and transliteration techniques negatively impacts classifier performance.  \nKeywords: Hate speech detection, transliteration, Amharic words, Latin script, single channel CNN, Multichannel, SVM  \n1. Introduction  \nThe use of social media platforms, such as Facebook and Twitter, has significantly increased user participation in communication, particularly in political, economic, and social domains [1] . Users actively engage in activities like political discussions, online commerce, and social matters, primarily through informal writing forms such as comments, posts, and chats [2] . This informal writing can be produced in native  \n59  \nJournal of Computational Science & Data Analytics © AASTU Press  \n60 Abebaw, [Z. et. al](Z. et. al).  \nlanguage scripts or transliterated scripts. Transliteration involves converting text from one script to another [3][4], such as transforming the Amharic term “በጣምጥሩ” (pronounced “bt’äm t’ru,” meaning “very good” in English) into its Latin script equivalent,“bettam ttiru.” The use of transliterated scripts on social media presents significant challenges for automated hate speech detection models. These models, typically trained on formal language scripts, often fail to detect hate speech written in transliterated forms, enabling hate speech to bypass detection mechanisms. For instance, while a hate speech comment such as ”destroy all blacks” in English or “ሁሉን ጥቁር አጥፋ” in Amharic might be detected by models trained on formal scripts, the transliterated version ”hulun ttiqur attfa” would likely evade detection. This gap in model capabilities highlights the urgent need to address the challenges posed by transli","cbCaicBVne8BlMYq","https://ap.wps.com/l/cbCaicBVne8BlMYq","pdf",161730,17,"English","# Introduction\n## Contributions and study organization\n# Literature review\n## Hate speech","[{\"question\":\"What research gap does the study address about Amharic hate speech detection?\",\"answer\":\"The study targets the lack of research on detecting hate speech from transliterated Amharic comments, especially those posted on social media, and evaluates how auto- vs manually transliterated data affects model performance.\"},{\"question\":\"Which classifiers are evaluated in the study?\",\"answer\":\"The study evaluates Support Vector Machine (SVM) and convolutional neural network (CNN) approaches, including single-channel and multichannel CNN models.\"},{\"question\":\"What is the key finding about using transliterated datasets together with the existing dataset?\",\"answer\":\"Combining transliterated datasets with the existing Amharic dataset does not improve classifier performance, likely due to inconsistencies in transliteration and dataset domain dependencies.\"}]","Hate Speech Detection from Transliterated Amharic Social Media Comments Using Machine Learning and Deep Learning Approaches | PDF",43]