DIAGNOSTIC ASSISTANCE FOR THE PREDICTION OF CARDIAC DISEASES WITH A NON-INVASIVE ESTIMATION OF HEMODYNAMIC PARAMETERS - Thesis

Aug 5, 2026

Upload by Aria

Machine learning is applied to support diagnostic assistance for cardiac diseases through non-invasive estimation of hemodynamic parameters. The work frames cardiac output measurement theory alongside arterial blood pressure and cardiac cycle concepts, then details data exploration, feature engineering, and feature selection strategies. A supervised learning pipeline is built with cross-validation, including waveform processing and selection of algorithms. Results cover signal processing using WFDB, model performance comparisons, and interpretation through a discussion that leads to conclusions and future work directions.

LanguageEnglish

CategoryResearch & Report

FormatPDF

Pages75

Reading Time3 h 9 min

AccessFree to download

Background

Problem Statement

Research Objectives

Aim

Objectives

Scope and Limitation

Theoretical Framework

Cardiac Output Measurement Theory

Arterial Blood Pressure

Cardiac Cycle

Exploration of Data

Feature Engineering

Feature Selection Techniques

Supervised Learning Algorithms

Cross-Validation

Related Work

Proposed Approach

Research Methodology

General Description

Data Acquisition

Data Filtering

Waveform Processing Processing (MGH/MF)

Feature Extraction and Algorithm Selection

Results

Signal Processing with WFDB

Machine Learning algorithms

Discussion and Conclusion

Future Work

References

DIAGNOSTIC ASSISTANCE FOR THE PREDICTION OF CARDIAC DISEASES WITH A NON-INVASIVE ESTIMATION OF HEMODYNAMIC PARAMETERS USING MACHINE LEARNING By Kamila Joseph Hernández Vanegas A THESIS Submitted in partial fulfillment of the requirements for the degree of MASTER OF SCIENCE In Electronics Engineering Universidad del Norte 2023 © Biomedical Signal Processing and Artificial Intelligence Laboratory (BSPAI)– Research Group This thesis has been approved in partial fulfillment of the requirements for the Degree of MASTER OF SCIENCE in In Electronics Engineering . Department of Electrical and Electronics Engineering Thesis Supervisor: Winston Spencer Percybrooks Bolivar, PhD. Program Coordinator: Mauricio Pardo Gonzales, PhD. Contenido 1 Background ...................................................................................................... 1 1.1 Problem Statement ..............................................................................5 2 Research Objectives .........................................................................................9 2.1 Aim .....................................................................................................9 2.2 Objectives ...........................................................................................9 2.3 Scope and Limitation ..........................................................................9 3 Theoretical Framework ..................................................................................11 3.1 Cardiac Output Measurement Theory...............................................11 3.2 Arterial Blood Pressure .....................................................................13 3.3 Cardiac Cycle .................................................................................... 14 3.4 Exploration of Data ........................................................................... 15 3.5 Feature Engineering .......................................................................... 15 3.6 Feature Selection Techniques ...........................................................15 3.7 Supervised Learning Algorithms ......................................................16 3.8 Cross-Validation ............................................................................... 17 4 Related Work.................................................................................................. 19 5 Proposed Approach ........................................................................................25 6 Research Methodology ...................................................................................28 6.1 General Description ..........................................................................28 6.2 Data Acquisition ...............................................................................29 6.3 Data Filtering ....................................................................................33 6.4 Waveform Processing Processing (MGH/MF) .................................36 6.5 Feature Extraction and Algorithm Selection.....................................39 7 Results ............................................................................................................47 7.1 Signal Processing with WFDB .........................................................47 7.2 Machine Learning algorithms ...........................................................51 8 Discussion and Conclusion ............................................................................55 8.1 Future Work ......................................................................................56 9 References ......................................................................................................57 List of Figures Fig.1 . 1. Characterization of channels with its corresponding signals, from the MIMIC II waveform visualizer.......................................

FAQs

What is the main goal of the thesis?

To provide diagnostic assistance for predicting cardiac diseases by estimating hemodynamic parameters using non-invasive data and machine learning.

Which data processing and modeling steps are included?

The methodology covers data acquisition and filtering, waveform processing (MGH/MF), feature extraction and algorithm selection, and cross-validation for supervised learning.

How are the results evaluated and discussed?

Results include signal processing with WFDB and performance of machine learning algorithms, followed by a discussion that concludes the work and outlines future directions.

Aug 5, 2026

Upload by Aria