Internet of Things and Machine Learning for Type I and Type II Diabetes : Use Cases.

Format:

Book

Author/Creator:

Dash, Sujata.

Contributor:

Pani, Subhendu Kumar.

Susilo, Willy.

Yung Bernard, Cheung Man.

Tse, Gary.

Language:

English

Subjects (All):

Diabetes.

Machine learning.

Physical Description:

1 online resource (450 pages)

Edition:

1st ed.

Place of Publication:

San Diego : Elsevier, 2024.

Summary:

This book explores the role of the Internet of Things (IoT) and machine learning in the diagnosis and management of Type I and Type II diabetes. It provides an in-depth analysis of intelligent systems for diabetes diagnosis, including rule-based machine learning techniques and deep learning models. The book covers various machine learning algorithms and their applications in predicting glucose levels, diagnosing diabetic retinopathy, and assessing diabetes-related risks. It aims to enhance healthcare through the integration of IoT and machine learning technologies, offering insights for researchers, practitioners, and healthcare professionals interested in the latest advancements in diabetes management and predictive analytics. Generated by AI.

Contents:

Front Cover

INTERNET OF THINGS AND MACHINE LEARNING FOR TYPE I AND TYPE II DIABETES

INTERNET OF THINGS AND MACHINE LEARNING FOR TYPE I AND TYPE II DIABETES: USE CASES

Copyright

Contents

Contributors

Preface

I - Diagnosis

1 - A systematic review on intelligent diagnosis of diabetes using rule-based machine learning techniques

1. Introduction

2. Literature search strategy

3. System overview

4. Dataset

5. Preprocessing methods

6. Algorithms for classification

6.1 Fuzzy system

6.2 SVM

6.3 Other algorithms

7. Application scenarios

8. Limitations and future directions

9. Conclusions

References

2 - Ensemble sparse intelligent mining techniques for diabetes diagnosis

2. Literature survey

3. Methodology

3.1 Data extraction

3.1.1 Dataset

3.2 Data preprocessing

3.2.1 Data cleaning and encoding

3.2.2 Data splitting

3.3 Model building

3.3.1 Libraries and modules used

3.4 Algorithms used

3.4.1 Existing models

3.4.1.1 Multilayer Perceptron (MLP)

3.4.1.2 Extra trees classifier

3.4.1.3 Linear Discriminant Analysis

3.4.1.4 Proposed model

4. Result analysis

4.1 Evaluation measures

4.1.1 Matthew's correlation coefficient

4.1.2 Log loss

4.1.3 Receiver Operator Characteristics

4.1.4 F1 score

4.2 Results obtained

4.2.1 Multilayer Perceptron

4.2.2 MCC mathematical calculations

4.2.3 Extra trees classifier

4.2.4 Mathematical calculations of ROC

4.2.5 True positive rate (TPR)

4.2.6 False Positive Rate (FPR)

4.2.7 Linear Discriminant Analysis

4.2.8 Stacking classifier

4.2.9 Mathematical calculation of F1 - score

4.3 Comparing algorithms

5. Conclusion and Future work

3 - Detection of diabetic retinopathy using Deep Neural networks

1.1 Diabetic retinopathy.

1.2 Prevalence of diabetic retinopathy in India

1.3 Traditional diagnosis of diabetic retinopathy

1.4 Machine learning

2. System analysis

2.1 Supervised learning

2.2 Deep learning

2.3 Convolutional neural networks

2.3.1 Image classification and recognition

2.4 Classifiers

2.4.1 Linear classifiers

2.5 Previous automation techniques for diagnosis of diabetic retinopathy

2.5.1 Comparison with present techniques used in the project

3. Modules

3.1 TensorFlow installation

3.2 Downloading the training data set

3.3 Training the CNN model

3.3.1 Bottlenecks

3.3.2 Training

3.4 Testing

4. System design

5. Implementation

5.1 Graph file and label file

5.2 Testing the trained model

6. Results

7. Conclusion and future scope

4 - An intelligent remote diagnostic approach for diabetes using machine learning techniques

1.1 Remote healthcare monitoring system

1.2 ECG and diabetes

1.3 Motivation

2. Intelligent remote diagnostic approach for diabetes

2.1 ECG signal acquisition and processing

2.1.1 Adaptive filtering

2.1.2 Wavelet denoising

2.1.3 Event detection and segmentation

2.2 Feature extraction

2.2.1 Time domain features

2.2.2 Frequency domain features

2.2.3 Time-frequency domain features

2.2.3.1 Short-time Fourier transform

2.2.4 Continuous wavelet transform

2.3 Classification and evaluation

3. Conclusion

5 - Diagnosis of diabetic retinopathy in retinal fundus images using machine learning and deep learning models

5Data availability

2. Dataset

3. Texture feature extraction

3.1 GLCM features

3.2 GLRLM features

3.3 Laws texture feature

4. Transform based feature extraction

4.1 Gabor transform

4.2 Radon transform

5. Dimensionality reduction.

6. Classification

6.1 k-NN classifier

6.2 SVM classifier

7. CNN-based deep learning algorithm for DR classification

7.1 Data augmentation

7.2 Proposed CNN architecture I

7.2.1 Convolutional layer

7.2.2 Activation function

7.2.3 Pooling layer

7.2.4 Dense and dropout layers

7.2.5 Output classification layer

7.3 Batch normalization

7.4 Optimization

7.5 Feature visualization

7.6 Performance analysis

8. Summary

6 - Diagnosis of diabetes mellitus using deep learning techniques and big data

2. Literature review

2.1 Motivation

3. Materials and methods

3.1 Proposed methodology

3.2 Data description

3.3 Data cleaning

3.4 Various challenges in the dataset

3.5 Inclusion and exclusion criteria of the patients

3.6 Preprocessing of data

3.6.1 Missing value handling technique

3.6.1.1 k-nearest neighbor (k-NN) imputation

3.6.2 Balancing of data

3.6.2.1 Synthetic minority oversampling technique

3.6.3 Feature extraction technique

3.6.3.1 Discrete wavelet transform (DWT)

3.6.4 Data normalization

3.7 Model development

3.7.1 Deep neural network (DNN)

3.7.2 Deep long-short-term memory (DLSTM)

4. Results and discussions

4.1 Simulation study

4.2 Performance measures

4.3 Results

4.4 Comparison with the winner list of WiDS Datathon 2021

5. Conclusion and future work

II - Glucose monitoring

7 - IoT and machine learning for management of diabetes mellitus

2. IoT and machine learning in general

3. Rationale of integrating IoT and machine learning in management of diabetes

4. IoT and machine learning in diabetes

4.1 IoT in diabetes management

4.1.1 Hand-held IoT-Based tablets

4.1.2 Wearable IoT-Based devices

4.1.3 Nanochips and sensors

4.1.4 Implants.

4.2 Machine learning tools for management of diabetes

4.2.1 Categories of ML learning processes

4.2.1.1 Supervised learning

4.2.1.2 Unsupervised learning

4.2.1.3 Reinforcement learning

4.2.1.4 Feature selection

4.2.2 Classification algorithms

4.2.2.1 SVM classifier

4.2.2.2 KNN classifier

4.2.2.3 Random Forest

4.2.2.4 ANN classifier

4.2.2.5 Naive Bayes classifier

5. Proposed framework and methodology

6. Future of IoT and machine learning

7. Conclusions

8 - Prediction of glucose concentration in type 1 diabetes patients based on machine learning techniques

2. Glucose management in type 1 diabetes

3. Machine learning in healthcare

4. Predicting glucose concentrations using machine learning

5. Linear regression

6. Support vector machines

7. Random forest models

8. Deep learning models

9. Conclusion

9 - ML-based PCA methods to diagnose statistical distribution of blood glucose levels of diabetic patients

2. Related algorithms

2.1 Principal component analysis

2.2 Kernel principal component analysis (KPCA)

2.3 Least squares vector machine

3. Prediction of fasting blood glucose level based on KPCA-LSSVM

4. Experimental methods

4.1 Original data source and preprocessing

4.2 Kernel principal component analysis

4.3 Least squares support vector machine modeling

4.4 Methodological analysis

5. Conclusions

Further reading

III - Prediction of complications and risk stratification

10 - Overview of new trends on deep learning models for diabetes risk prediction

1.1 Abstraction in multiple layers

1.2 Larger datasets are beneficial for training

1.3 Feature extraction using automated means

1.4 Managing data from diverse sources

2. Overview of DL.

3. The identification of diabetes mellitus

4. Management of blood sugar

5. Complications and their diagnosis

6. An overview of DL methods in a Nutshell

7. Discussion

7.1 Constraints and obstacles in the way

7.2 Possibilities and work in the future

8. Conclusion

11 - Clinical applications of deep learning in diabetes and its enhancements with future predictions

2. Artificial intelligence

3. Diagnosis of diabetes mellitus

4. Diabetes-related complications

4.1 Retinopathy

4.2 Diabetic foot ulcer

4.3 Diabetic neuropathy

5. Glucose measurement and prediction

5.1 Continuous glucose monitoring

5.2 Hypoglycemic episodes

5.3 Glucose prediction

6. Conclusion/future aspect

12 - Exploring machine learning techniques for feature extraction and classification of diabetes related medical da ...

2.1 Diabetes classification on clinical datasets

2.2 Diabetes detection and classification on retinal image data

3. Diabetes datasets

3.1 PIMA Indians dataset

3.2 Asia Pacific tele-ophthalmology society (APTOS) dataset

4. Preprocessing

4.1 Handling missing values

4.2 Filtering outliers

4.3 Feature extraction

5. Classification techniques

5.1 Methods used with numerical datasets

5.1.1 Support vector machine

5.1.2 Logistic regression

5.1.3 Decision tree

5.1.4 Random forest

5.1.5 AdaBoost

5.1.6 XGBoost

5.1.7 K-nearest neighbors

5.1.8 Artificial neural network

5.1.9 Deep neural network

5.1.10 Recurrent neural network

5.1.11 Long short-term memory

5.2 Methods used with diabetes image datasets

5.2.1 Convolutional neural networks

5.2.2 Deep convolutional neural network

5.2.3 A residual network (ResNet)

5.2.4 Distributed deep learning.

6. Comparative analysis and discussion.

Notes:

Description based on publisher supplied metadata and other sources.

Part of the metadata in this record was created by AI, based on the text of the resource.

ISBN:

9780323956932

0323956939

OCLC:

1446129500