[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"doc-detail-123408-en":3,"doc-seo-123408-105":30,"detail-sidebar-cat-0-en-105":95},{"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},123408,13056703019662,"Evangeline","https://ap-avatar.wpscdn.com/avatar/be000253a8e92610077?_k=1778726343310543188",8,"Research & Report","MACHINE LEARNING-BASED ANALYSIS OF ELECTROCARDIOGRAMS FOR ANOMALY DETECTION - Bachelor of Science Thesis","Electrocardiograms (ECGs) have long supported cardiovascular disease study and diagnosis, motivating automated interpretation of ECG signals. This Bachelor of Science thesis experimentally evaluates multiple machine learning models for anomaly detection in ECG data, using an MIT–Beth Israel Hospital dataset created in the 1970s. The dataset is normalized, filtered, segmented into ten-second samples, and split consistently into training and testing sets. Feature extraction is varied via PCA with different kernels and via FFT followed by linear PCA, then the tested models’ performance is measured with five metrics. Results show that simpler models train much faster than state-of-the-art deep networks, while accuracy could be improved by deepening and tailoring preprocessing for specific models like the 1DCNN.","Tommi Salonen  \nMACHINE LEARNING-BASED ANALYSIS OF ELECTROCARDIOGRAMS FOR ANOMALY DETECTION  \nBachelor of Science Thesis  \nFaculty of Information Technology and Communication Sciences Examiner: Dr. Fahad Sohrab  \nMay 2025  \ni  \nABSTRACT  \nTommi Salonen: Machine Learning-Based Analysis of Electrocardiograms for Anomaly Detection Bachelor of Science Thesis  \nTampere University  \nBachelor of Science (Tech), Computing and Electrical Engineering May 2025  \nElectrocardiograms (ECG) have been used for decades in the study and diagnosis of cardiovascular diseases. Efforts to automate the interpretation of ECG signals have existed for nearly as long. The development of machine learning algorithms has enabled the creation of increasingly accurate and efficient diagnostic systems. This thesis experimentally investigates the ability of various machine learning models to detect anomalies in ECG data. The dataset used in this study was created as a collaboration between the Massachusetts Institute of Technology and Beth Israel Hospital in the 1970s. To ensure comparability, the data is preprocessed and split ina consistent manner across all models. The preprocessing steps include normalization, filtering, segmentation into ten-second samples, and dividing into training and testing sets. In addition, the thesis explores how different feature extraction techniques influence classification performance. The final preprocessing step is varied to assess its effect. The tested methods include Principal Component Analysis (PCA) with various kernel functions, as well as a combination of Fast Fourier Transform (FFT) followed by linear PCA. The machine learning models evaluated in this study are: K-Nearest Neighbors (KNN), Support Vector Machine (SVM), Random Forest (RF), and a 1-Dimensional Convolutional Neural Network (1DCNN) . The models’ performance is assessed using five different metrics to provide a comprehensive evaluation. The goal of the thesis is to compare the accuracy of different models and preprocessing combinations, and to identify which approach performs best using relatively simple preprocessing techniques. While the results fall short when compared to state-of-the-art deep neural networks, the simplicity of the models used makes them significantly faster to train. For example, the accuracy of the 1DCNN could be improved substantially by deepening the network and optimizing the preprocessing pipeline specifically for that model.  \nKeywords: machine learning, electrocardiogram, arrhythmia detection, biomedical signal processing, MIT-BIH arrhythmia database  \nThe originality of this thesis has been checked using the Turnitin OriginalityCheck service.  \nii  \nTIIVISTELMÄ  \nTommi Salonen: Koneoppimiseen perustuva elektrokardiogrammidatan analysointi poikkeavuuksien löytämiseksi  \nKandidaatintyö Tampereen yliopisto  \nTieto-ja sähkötekniikan kandidaattiohjelma Toukokuu 2025  \nElektrokardiogrammeja (EKG) on käytetty sydän-ja verisuonitautien tutkimiseen jo vuosikymmeniä . EKG:n tulkinnan automatisointia on yritetty kehittää lähes yhtä kauan. Koneoppimisalgoritmien kehitys on mahdollistanut entistä tarkempien ja tehokkaampien järjestelmien kehityksen. Tässä tutkielmassa tutkitaan kokeellisesti erilaisten koneoppimismallien kykyä tunnistaa poikkeavuuksia elektrokardiogrammidatasta. Tutkielmassa käytetty data on peräisin Massachusetts Institute of Technologyn ja Beth Israel Hospitalin yhteistyössä tuottamasta datasarjasta. Data esikäsitellään ja jaetaan samalla tavalla kaikkien algoritmien käyttöön, jotta tulokset ovat vertailukelpoisia. Datan esikäsittelyyn kuuluu datan normalisointi, suodatus, pilkkominen kymmenen sekunninnäyttesiin sekä jakaminen testaus-ja harjoitusjoukkoihin. Näiden toimenpiteiden jälkeen datalle tehdään pääkomponenttianalyysi. Tutkielmassa selvitetään myös erilaisten esikäsittelytekniikoiden vaikutusta luokittelutuloksiin. Datan esikäsittelyssä muokataan viimeistä vaihetta. Tutkittavat menetelmät ovat pääkomponenttianalyy","cbCaijKuGLfkAUyX","https://ap.wps.com/l/cbCaijKuGLfkAUyX","pdf",388896,1,29,"English","en",105,"# Abstract\n## Methods and preprocessing\n## Feature extraction and model evaluation\n# Keywords\n# Use of AI in thesis\n## AI tools and purposes\n## Accountability statement\n# Preface","[{\"question\":\"What is the main goal of the thesis?\",\"answer\":\"To compare multiple machine learning models and preprocessing combinations for detecting anomalies in ECG data, and identify which approach performs best using relatively simple preprocessing techniques.\"},{\"question\":\"How is the ECG dataset prepared before training and testing?\",\"answer\":\"The data is normalized, filtered, segmented into ten-second samples, and split into consistent training and testing sets across all models.\"},{\"question\":\"Which feature extraction techniques and models are evaluated?\",\"answer\":\"Feature extraction uses PCA with various kernels, and also FFT followed by linear PCA. The evaluated models include KNN, SVM, Random Forest, and a 1D convolutional neural network (1DCNN).\"},{\"question\":\"What does the thesis conclude about the trade-off between accuracy and training speed?\",\"answer\":\"The results are weaker than state-of-the-art deep neural networks, but the simpler models are significantly faster to train; accuracy can improve by deepening the 1DCNN and optimizing its preprocessing pipeline.\"}]","MACHINE LEARNING-BASED ANALYSIS OF ELECTROCARDIOGRAMS FOR ANOMALY DETECTION - Bachelor of Science Thesis | PDF",1785816324,73,{"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":90,"head_meta":92,"extra_data":94,"updated_unix":28},"machine-learning-based-analysis-of-electrocardiograms-for-anomaly-detection-bachelor-of-science-thesis","",{"@graph":36,"@context":89},[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/machine-learning-based-analysis-of-electrocardiograms-for-anomaly-detection-bachelor-of-science-thesis/123408/",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","2026-08-04",true,{"@type":65,"interactionType":66,"userInteractionCount":4},"InteractionCounter",{"@type":67},"ViewAction",{"@type":69,"mainEntity":70},"FAQPage",[71,77,81,85],{"name":72,"@type":73,"acceptedAnswer":74},"What is the main goal of the thesis?","Question",{"text":75,"@type":76},"To compare multiple machine learning models and preprocessing combinations for detecting anomalies in ECG data, and identify which approach performs best using relatively simple preprocessing techniques.","Answer",{"name":78,"@type":73,"acceptedAnswer":79},"How is the ECG dataset prepared before training and testing?",{"text":80,"@type":76},"The data is normalized, filtered, segmented into ten-second samples, and split into consistent training and testing sets across all models.",{"name":82,"@type":73,"acceptedAnswer":83},"Which feature extraction techniques and models are evaluated?",{"text":84,"@type":76},"Feature extraction uses PCA with various kernels, and also FFT followed by linear PCA. The evaluated models include KNN, SVM, Random Forest, and a 1D convolutional neural network (1DCNN).",{"name":86,"@type":73,"acceptedAnswer":87},"What does the thesis conclude about the trade-off between accuracy and training speed?",{"text":88,"@type":76},"The results are weaker than state-of-the-art deep neural networks, but the simpler models are significantly faster to train; accuracy can improve by deepening the 1DCNN and optimizing its preprocessing pipeline.","https://schema.org",{"og:url":52,"og:type":91,"og:title":13,"og:site_name":59,"og:description":14},"article",{"robots":93,"canonical":52},"index,follow",{"doc_id":7,"site_id":24},{"code":4,"msg":5,"data":96},[97,101,105,109,114,119,124,127,132,135,139],{"id":20,"doc_module":4,"doc_module_name":46,"category_name":98,"show_sort_weight":99,"slug":100},"Story & Novel",90,"story-novel",{"id":47,"doc_module":4,"doc_module_name":46,"category_name":102,"show_sort_weight":103,"slug":104},"Literature",80,"literature",{"id":53,"doc_module":4,"doc_module_name":46,"category_name":106,"show_sort_weight":107,"slug":108},"Exam",70,"exam",{"id":110,"doc_module":4,"doc_module_name":46,"category_name":111,"show_sort_weight":112,"slug":113},5,"Comic",60,"comic",{"id":115,"doc_module":4,"doc_module_name":46,"category_name":116,"show_sort_weight":117,"slug":118},6,"Technology",50,"technology",{"id":120,"doc_module":4,"doc_module_name":46,"category_name":121,"show_sort_weight":122,"slug":123},7,"Healthcare",40,"healthcare",{"id":11,"doc_module":4,"doc_module_name":46,"category_name":12,"show_sort_weight":125,"slug":126},30,"research-report",{"id":128,"doc_module":4,"doc_module_name":46,"category_name":129,"show_sort_weight":130,"slug":131},9,"Religion & Spirituality",20,"religion-spirituality",{"id":130,"doc_module":4,"doc_module_name":46,"category_name":133,"show_sort_weight":130,"slug":134},"World Cup","world-cup",{"id":136,"doc_module":4,"doc_module_name":46,"category_name":137,"show_sort_weight":136,"slug":138},10,"Lifestyle","lifestyle",{"id":140,"doc_module":4,"doc_module_name":46,"category_name":141,"show_sort_weight":110,"slug":142},19,"General","general"]