{"componentChunkName":"component---src-templates-project-en-js","path":"/en/projects/arrhythmia-ecg-recognition","webpackCompilationHash":"616a92a4504883dda87b","result":{"pageContext":{"isCreatedByStatefulCreatePages":false,"id":"arrhythmia-ecg-recognition","lang":"en","content":{"id":"arrhythmia-ecg-recognition","header":{"title":"Arrhythmia ECG Recognition Project","subtitle":"Arrhythmia ECG recognition project developed with discrete wavelet transform and machine learning tools.","links":[{"title":"Github","url":"https://github.com/davikawasaki/arrhythmia-ecg-analysis-ai","icon":"socialMediaGithubCircleIcon"}],"date":{"start":"1494853200","end":"1508418000"}},"about":{"categories":[{"title":"Software Work","abbr":"SW"},{"title":"Machine Learning","abbr":"ML"},{"title":"Data Science","abbr":"DS"},{"title":"Digital Signal Processing","abbr":"DSP"}],"text":["Artificial intelligence project developed through undergraduate course focused on cardiac arrhythmia classification using supervised machine learning. This is done through extraction of mixed types of heart waves characteristics from electrocardiograms (ECG) using discrete wavelet transform to filter the signal. The exported data was used in a list machine learning algorithms classifying the exported characteristics with classes/true labels. The goal, in the end, was to classify at least two arrhythmia through some extracted characteristics using the softwares Weka and MATLAB.","With the evaluation test (split instances and confusion matrix) results for each software, the ECG arrhythmia extraction and analysis were well evaluated on testing for the Support-vector machine (97,90%) and k-nearest neighbors (91,49%) algorithms, while for the Naive Bayes (74,47%) and BoostedTrees (21,30%) algorithms the results were the worst ones.","Roles: Data Engineer, Developer.","Tech stack: MATLAB R2017b, Weka 3.8.1."]},"gallery":{"title":"Project Highlights","photos":[]},"contact":{"title":"Contact for partnerships or proposals","links":[{"type":"email","svg":"socialMediaEmailCircleIcon","url":"mailto:davishinjik@gmail.com"},{"type":"linkedin","svg":"socialMediaLinkedinCircleIcon","url":"https://www.linkedin.com/in/davikawasaki/"}]},"footer":{"text":"© Since 2017. Made with ♥ in Denmark."}}}}}