Development of machine-learning-based natural language processing to detect concept labels in clinical narratives

Aug 5, 2026

Upload by Miles

Clinical care generates abundant textual data, yet narrative notes written by physicians remain underused because they are largely unstructured or only semi-structured. This work leverages machine-learning and deep learning to capture latent representations of clinical narratives, while using natural language processing during preprocessing to translate raw text into a lower-dimensional feature space. The approach aims to improve performance on clinical notes and enable reliable detection of concept labels, supporting more effective and personalized health services from routine narrative records.

LanguageEnglish

CategoryHealthcare

FormatPDF

Pages151

Reading Time6 h 21 min

AccessFree to download

Résumé

Données cliniques et problème des notes narratives

Apprentissage en profondeur et représentation latente

Rôle du NLP au prétraitement

Objectif : détection des concepts

Development of machine-learning-based natural language processing to detect concept labels in clinical narratives by Thanh-Dung LE MANUSCRIPT-BASED THESIS PRESENTED TO ÉCOLE DE TECHNOLOGIE SUPÉRIEURE IN PARTIAL FULFILLMENT FOR THE DEGREE OF DOCTOR OF PHILOSOPHY Ph.D. MONTREAL, SEPTEMBER 12, 2023 ÉCOLE DE TECHNOLOGIE SUPÉRIEURE UNIVERSITÉ DU QUÉBEC Thanh-Dung Le, 2023 This Creative Commons license allows readers to download this work and share it with others as long as the author is credited. The content of this work cannot be modiﬁed in any way or used commercially. BOARD OF EXAMINERS THIS THESIS HAS BEEN EVALUATED BY THE FOLLOWING BOARD OF EXAMINERS Mrs. Rita Noumeir, Thesis supervisor Department of Electrical Engineering, École de Technologie Supérieure Mr. Philippe Jouvet, Thesis Co-Supervisor Pediatric Intensivist-Ste. Justine Hospital Montréal, Université de Montréal Mr. Mohamed Cheriet, Chair, Board of Examiners Department of Electrical Engineering, École de Technologie Supérieure Mrs. Rachel Bouserhal, Member of the Jury Department of Electrical Engineering, École de Technologie Supérieure Mrs. Sabine Bergler, External Examiner Department of Computer Science and Software Engineering, Concordia University THIS THESIS WAS PRESENTED AND DEFENDED IN THE PRESENCE OF A BOARD OF EXAMINERS AND THE PUBLIC ON AUGUST 14, 2023 AT ÉCOLE DE TECHNOLOGIE SUPÉRIEURE ACKNOWLEDGEMENTS My journey toward a Ph.D. has been challenging yet rewarding, and I am grateful for the unwavering support and encouragement I received from my supervisors, colleagues, friends, and family. While it’s diﬃcult to express my sincere gratitude to them in words, I would like to take this opportunity to acknowledge their contributions. First and foremost, I express my deepest gratitude to my principal supervisor, Professor Rita Noumeir, for her continuous guidance, motivation, and inspiration throughout my Ph.D. studies. Her invaluable advice and discussions during weekly meetings have positively impacted my research and career trajectory. I am also grateful for her support during the COVID-19 pandemic, where her instructions and supervision helped me adapt to remote working tools and maintain my research progress. I also thank my co-supervisors, Professor Phillipe Jouvet, for his valuable comments and feedback on my research manuscripts. I would also like to thank my Ph.D. assessment jury for their constructive feedback and support. I thank Fonds de Recherche du Québec Nature et Technologies for awarding me the merit scholarship that enabled me to pursue my Ph.D. project. I also want to thank Pervasive and Smart Wireless Applications for the Digital Economy (PERSWADE) for their support during my research at The Énergie, Matériaux et Télécommunications center, Institut National de la Recherche Scientiﬁque (INRS) . Additionally, I would like to thank the anonymous reviewers who provided valuable feedback on my research papers and helped improve my work’s quality. I acknowledge my former supervisors and colleagues at the NECHPY-Lab, INRS: Professor Long Le, Dr. Ha Nguyen Vu, Dr. Hoang Duc Tuong, Dr. Tran Duy Thinh, Dr. Nguyen Ti Ti, Dr. Nguyen Minh Tri, Vu Huy Hoang, Phan Thanh Tung, Dr. Nguyen Minh Dat. Furthermore, I express gratitude to my colleagues at LATIS, ETS: Dr. Georges Matar, Jihad El Tannoury, Gloria Huong, Oussema, Haythem, Mario, Asheok, Wajahat, Khalil, Touﬁk, Clara, Emily, and many others for their support, encouragement, and the memories we have shared. VI Finally, I am indebted to my parents (Lam-Bup, Luyen-La), sisters (Hong Diem, Ngoc Trinh, Le Huyen), brothers (Quoc Vuong, Xuan Son), niece (Hong Vy), nephews (Gia Bao, Thien Phuc) for their unconditional love, support, and sacriﬁces. Their emotional and ﬁnancial support have been invaluable during my Ph.D. journey. Special thanks to my wife and our little princess Nha Lam Cecilia, who has been with me every step of the way, provi

FAQs

Pourquoi les notes cliniques narratives sont-elles peu utilisées en pratique ?

Elles sont majoritairement sous un format non structuré ou semi-structuré, ce qui complique leur exploitation directe malgré leur disponibilité dans les entrepôts de données.

Comment l’apprentissage en profondeur aide-t-il dans ce contexte ?

Il permet de capturer efficacement les représentations latentes des récits cliniques, grâce à sa puissance de calcul et sa capacité à apprendre des patterns complexes.

Quel est le rôle du NLP dans la méthode proposée ?

Le NLP est utilisé au prétraitement pour mapper les mots des données textuelles non structurées vers un espace de dimension inférieure, facilitant ensuite la tâche de détection de concepts.

Aug 5, 2026

Upload by Miles