Bio-Inspired Strategies for Modeling and Detection in Diabetes Mellitus Treatment
Bio-Inspired Strategies for Modeling and Detection in Diabetes Mellitus Treatment focuses on bio-inspired techniques such as modelling to generate control algorithms for the treatment of diabetes mellitus. The book addresses the identification of diabetes mellitus using a high-order recurrent neural...
| Autor principal: | |
|---|---|
| Otros Autores: | , , |
| Formato: | Libro electrónico |
| Idioma: | Inglés |
| Publicado: |
San Diego :
Elsevier Science & Technology
2024.
|
| Edición: | 1st ed |
| Materias: | |
| Ver en Biblioteca Universitat Ramon Llull: | https://discovery.url.edu/permalink/34CSUC_URL/1im36ta/alma991009811836106719 |
Tabla de Contenidos:
- Front Cover
- Bio-Inspired Strategies for Modeling and Detection in Diabetes Mellitus Treatment
- Copyright
- Dedication
- Contents
- List of figures
- List of tables
- Biographies
- Preface
- 1 Introduction
- 1.1 Diabetes Mellitus from a medical point of view
- 1.2 Current diagnosis of Diabetes Mellitus and glucose intolerance
- 1.2.1 Oral glucose tolerance test
- 1.3 Current treatment of Diabetes Mellitus
- 1.4 Prediction of Diabetes Mellitus and its complications
- 1.5 Intelligent strategies inspired in biology and their applications in engineering and medicine
- 1.5.1 Main branches of Computational Intelligence
- 1.5.1.1 Evolutionary computation
- 1.5.1.2 Swarm intelligence
- 1.5.1.3 Neural network
- 1.5.1.4 Fuzzy logic
- 1.5.1.5 Artificial immune system
- 1.5.1.6 Medical applications using bio-inspired strategies
- References
- 2 Problem statement
- 2.1 State-of-the-art in modeling, identification, and detection of diabetes mellitus using bio-inspired strategies
- 2.1.1 Parametric estimation
- 2.1.2 Neural identification
- 2.1.3 Prediction
- 2.1.4 Detection
- 2.2 Compartmental models Sorensen and Dalla Man
- 2.2.1 Sorensen model
- 2.2.2 Glucose subsystem
- 2.2.3 Insulin subsystem
- 2.2.4 Metabolic rates and dynamics of glucagon
- 2.2.5 Dalla Man model
- 2.3 Serial data
- 2.3.1 Continuous glucose monitoring system data
- 2.3.2 Tolerance tests to oral intake data
- References
- 3 Mathematical preliminaries
- 3.1 Evolutionary algorithms
- 3.1.1 Evonorm
- 3.1.2 Differential evolution
- 3.2 Particle swarm-based algorithms
- 3.2.1 Particle swarm optimization
- 3.2.2 Ant colony optimization
- 3.3 Neural networks
- 3.3.1 Multilayer perceptron neural network
- 3.3.2 Discrete-time high order neural networks
- 3.3.2.1 RHONN training algorithm
- 3.4 Deep neural networks.
- 3.4.1 Convolutional neural network
- 3.4.2 Long short-term memory recurrent neural network
- 3.4.3 Bidirectional LSTM
- 3.4.4 LSTM fully convolutional networks
- 3.4.5 Residual deep networks
- References
- 4 Parameter estimation for glucose-insulin dynamics
- 4.1 Affine system
- 4.1.1 Affine system of the Dalla Man model
- 4.1.2 Affine system of Sorensen model
- 4.2 Evolutionary optimization algorithms for parameter estimation
- 4.2.1 Methodology for parametric estimation
- 4.3 Parametric estimation results in compartmental models
- 4.3.1 Sorensen parametric estimation results
- 4.3.1.1 Results obtained by Evonorm
- 4.3.1.2 Results obtained by Differential Evolution
- 4.3.1.3 Results obtained by PSO
- 4.3.1.4 Results obtained by ACO
- 4.3.2 Dalla Man parametric estimation results
- 4.3.2.1 Results obtained by Evonorm
- 4.3.2.2 Results obtained by DE
- 4.3.2.3 Results obtained by PSO
- 4.3.2.4 Results obtained by ACO
- References
- 5 Neural model for glucose-insulin dynamics
- 5.1 Identification
- 5.2 Identification with artificial neural networks
- 5.3 System description
- 5.3.1 Identification with artificial neural networks offline
- 5.3.1.1 Neuronal network configuration using CGM
- 5.3.1.2 LSTM identification results
- 5.3.1.3 BiLSTM results
- 5.3.1.4 MLP results
- 5.3.2 Identification with an artificial neural network online
- 5.3.2.1 RHONN identification structure
- 5.3.3 RHONN results
- References
- 6 Multistep predictor applied to T1DM patients
- 6.1 Prediction
- 6.1.1 Multistep ahead prediction strategies
- 6.1.2 Recursive strategy
- 6.1.3 Direct strategy
- 6.1.4 Neuronal network configuration using CGM data
- 6.2 Prediction results evaluation criteria
- 6.3 Results
- 6.3.1 Neural models with offline training
- 6.3.1.1 MLP results
- 6.3.1.2 LSTM results
- 6.3.1.3 BiLSTM results.
- 6.3.1.4 Results discussion
- 6.3.2 Neural models with online training
- 6.3.2.1 RHONN configuration for MSA prediction
- 6.3.3 RHONN results
- References
- 7 Classification and detection of diabetes mellitus and impaired glucose tolerance
- 7.1 Classification
- 7.1.1 Time series classification
- 7.2 K-means
- 7.3 Evaluation criteria
- 7.4 Results
- 7.4.1 MLP results
- 7.4.1.1 Virtual patient classification results
- 7.4.1.2 Real patient classification results
- 7.4.2 CNN results
- 7.4.2.1 Virtual patient classification results
- 7.4.2.2 Real patient classification results
- 7.4.3 LSTM results
- 7.4.3.1 Virtual patient classification results
- 7.4.3.2 Real patient classification results
- 7.4.4 LSTM-FCN results
- 7.4.4.1 Virtual patient classification results
- 7.4.4.2 Real patient classification results
- 7.4.5 ResNet results
- 7.4.5.1 Virtual patient classification results
- 7.4.5.2 Real patient classification results
- 7.4.6 K-Means results
- 7.4.6.1 Virtual patient classification results
- 7.4.6.2 Real patient classification results
- 7.5 Results discussion
- References
- 8 Conclusions and future work
- 8.1 Conclusions of Chapter 1
- 8.2 Conclusions of Chapter 2
- 8.3 Conclusions of Chapter 3
- 8.4 Conclusions of Chapter 4
- 8.5 Conclusions of Chapter 5
- 8.6 Conclusions of Chapter 6
- 8.7 Conclusions of Chapter 7
- 8.8 General conclusions
- A Model parameters
- A.1 Dalla Man nominal parameter values
- A.2 Sorensen nominal parameter values
- Index
- Back Cover.
