Please use this identifier to cite or link to this item: http://dspace.dtu.ac.in:8080/jspui/handle/repository/20636
Full metadata record
DC FieldValueLanguage
dc.contributor.authorDEEPANSHU-
dc.date.accessioned2024-08-05T07:07:01Z-
dc.date.available2024-08-05T07:07:01Z-
dc.date.issued2023-06-
dc.identifier.urihttp://dspace.dtu.ac.in:8080/jspui/handle/repository/20636-
dc.description.abstractThe goal of this thesis is to evaluate how different machine learning algorithms perform in detecting diabetes using the Pima dataset. Three models are examined for their accuracy: Random Forest (72%), Long-Short-Term-Memory Networks (LSTMs) (75.32%), & Extreme-Learning-Machine (ELM) (77.70%). The findings demonstrate that the ensemble model containing ELM, LSTM, and RF outperforms the individual models with an accuracy of 93%. To strengthen the ensemble model further, a unique strategy is proposed, comprising a mix of ELM and LSTM models together with the addition of a Random Forest classifier. Cross-validation is utilised to test the suggested model, indicating its improved performance compared to the individual models, with a mean AUC of 0.93. Subsequently, the model is trained on the complete dataset and achieves an accuracy of 93% and an AUC of 0.89 on the testing set. These results show that the proposed approach efficiently diagnoses diabetes and offers promise for assisting clinical decision-making.en_US
dc.language.isoenen_US
dc.relation.ispartofseriesTD-7042;-
dc.subjectDIABETES DIAGNOSIS ACCURACYen_US
dc.subjectENSEMBLE LEARNING APPROACHen_US
dc.subjectLSTMsen_US
dc.titleENHANCING DIABETES DIAGNOSIS ACCURACY WITH AN ENSEMBLE LEARNING APPROACHen_US
dc.typeThesisen_US
Appears in Collections:M.E./M.Tech. Information Technology

Files in This Item:
File Description SizeFormat 
DEEPANSHU M.Tech.pdf1.07 MBAdobe PDFView/Open


Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.