Biomedical and Other Applications of Soft Computing

Description

This book presents innovative intelligent techniques, with an emphasis on their biomedical applications. Although many medical doctors are willing to share their knowledge – e.g. by incorporating it in computer-based advisory systems that can benefit other doctors – this knowledge is often expressed using imprecise (fuzzy) words from natural language such as “small,” which are difficult for computers to process. Accordingly, we need fuzzy techniques to handle such words. It is also desirable to extract general recommendations from the records of medical doctors’ decisions – by using machine learning techniques such as neural networks. The book describes state-of-the-art fuzzy, neural, and other techniques, especially those that are now being used, or potentially could be used, in biomedical applications. Accordingly, it will benefit all researchers and students interested in the latest developments, as well as practitioners who want to learn about new techniques. Preface Contents Question-Answering System over Knowledge Graphs Using Analogical-Problem-Solving Approach 1 Introduction 2 Background 2.1 RDF Knowledge Graphs and RDF Query Language 2.2 Dependency Parser 3 LingTeQA: An Analogical-Problem-Solving QA System 3.1 LingTeQA: Representing Questions 3.2 LingTeQA: Generating Templates 4 LingTeQA: Answering Questions 4.1 LingTeQA: Answering Question Containing Linguistic Terms 4.2 LingTeQA: Answering Questions by Generating Linguistic Summaries 4.3 LingTeQA: Defining Linguistic Terms with a User-friendly Web Interface 5 Related Work 5.1 Question Answering 5.2 Linguistic Summarization of Numeric Data 6 Result and Conclusion References Fuzzy Transform on 1-D Manifolds 1 Introduction 2 Preliminaries 2.1 Fuzzy Partition 2.2 Topological Manifold 2.3 Properties of Manifold upper MM 3 Fuzzy Partition of a Manifold 4 Hilbert Space on a Manifold 4.1 Subspace upper L 2 Superscript m Baseline left parenthesis ModifyingAbove upper A With quotation dash Subscript i Baseline right parenthesisL2m(overlineAi) 5 upper F Superscript mFm-transform 5.1 upper F Superscript 0F0-transform and its Inverse 6 Conclusion References A Systematic Review of Privacy-Preserving Blockchain in e-Medicine 1 Introduction 2 Blockchain and the Electronic Health Records Model 2.1 Conventional Electronic Health Records Model 2.2 Blockchain-Based Privacy-Preserving Models in Electric Health Records 3 Privacy-Preserving Identity Management Systems and Platforms in Blockchain 4 Future Research Directions 5 Conclusions References Why Rectified Linear Neurons: Two Convexity-Related Explanations 1 Formulation of the Problem 2 Why Convexity 3 First Convexity-Related Explanation 4 Second Convexity-Related Explanation References How to Work? How to Study? Shall We Cram for the Exams? and How Is This Related to Life on Earth? 1 When to Switch Activities: Formulation of the Problem 2 Let Us Formulate This Problem in Precise Terms 3 Analysis of the Problem 4 Resulting Recommendations 5 How Is This Related to Life on Earth? References Why Quantum Techniques Are a Good First Approximation to Social and Economic Phenomena, and What Next 1 Formulation of the Problem 2 Where Can Such an Explanation Come from: General Analysis 3 First Explanation: Quantum Formulas Provide a Good Description for Many Phenomena in General 4 Second Explanation: Quantum Formulas Are the Computationally Fastest Way to Describe Nonlinear Phenomena 5 Third Explanation: Quantum Physics Has Many Solved Problems 6 Beyond Quantum References How the Pavement's Lifetime Depends on the Stress Level and on the Dry Density: An Explanation of Empirical Formulas 1 First Problem: Dependence on Stress 2 Analysis of the Problem 3 Invariance Requirement 4 Resulting Explanation 5 Second Problem: Dependence on Dry Density References Freedom of Will, Physics, and Human Intelligence: An Idea 1 Three Fundamental Challenges 2 How We Can Solve These Challenges: An Idea 3 Let Us Summarize Our Findings References Why Normalized Difference Vegetation Index (NDVI)? 1 Formulation of the Problem 2 Towards an Explanation: General Analysis of the Problem 3 First Result: Characterizing All Natural Functions of Two Variables 4 Scale Invariance References Binary Image Classification Using Convolutional Neural Network for V2V Communication Systems 1 Introduction 2 Related Work 2.1 Traditional Approaches 2.2 Deep Learning Approaches 3 Preliminaries of VOCC Systems 4 Proposed Algorithm 5 Experiment Analyses 5.1 Dataset 5.2 Evaluation metrics 5.3 Result 6 Conclusion References Topic Model—Machine Learning Classifier Integrations on Geocoded Twitter Data 1 Introduction 2 Related Literature 3 Data 4 Methods 4.1 Topic Models Formulation 4.2 From Topic Models to Machine Learning Classifiers 4.3 Artificial Neural Networks 5 Results 5.1 Topic Models 5.2 Machine Learning Classifiers 5.3 Artificial Neural Networks 6 Conclusion 7 Appendix References Shop Product Tracking and Early Fire Detection Using Edge Devices 1 Introduction 2 Related Work 2.1 Real World Applications 2.2 Technologies 3 Method 3.1 Architecture 3.2 Product Recognition Module 3.3 Fire Detection Module 3.4 Product Tracking Module 4 Experiments 4.1 Datasets 4.2 Choosing Algorithms 4.3 Application Testing 5 Conclusions References SDNs Delay Prediction Using Machine Learning Algorithms 1 Introduction 2 Related Work 2.1 Machine Learning Algorithms 2.2 Evaluation Metrics 3 Method 3.1 Data Collection 3.2 Data Preprocessing 3.3 Training and Evaluation 4 Experiments 4.1 Datasets and Data Preprocessing 4.2 Comparison Between Different Approaches 5 Conclusions References A Linear Neural Network Approach for Solving Partial Differential Equations on Porous Domains 1 Introduction 2 The Proposed Radial Basis Function Network Technique 2.1 Integrated RBF Network Discretisation for Extended Domain 3 Numerical Solutions 3.1 Elliptic Partial Differential Equation 3.2 Parabolic Partial Differential Equation 4 Concluding Remarks References Accuracy Measures and the Convexity of ROC Curves for Binary Classification Problems 1 Introduction 2 Convexity of the ROC Curves 3 Inequalities Relating Different Measures of Accuracy References Stochastic Simulations of Airborne Particles in a Fibre Matrix 1 Introduction 2 Langevin Equation 3 SDE Solver 3.1 Euler Method 3.2 Milstein Method 4 Langevin Equation for Particles' Zig-Zac Motion 5 Conclusion References Disease Diagnosis Based on Symptoms Description 1 Introduction 2 Overview of Related Works 3 Methodology 3.1 Background 3.2 Datasets, Input Pre-processing, Model Selection and Training 3.3 The Proposed Architecture for Diagnosis Based Symptom Descriptions 4 Results 4.1 Evaluation and Comparison with Prediction Performance of Bi-RNN 4.2 Practical Testing 5 Conclusion and Future Works References Chest X-Ray Image Analysis with ResNet50, SMOTE and SafeSMOTE 1 Introduction 2 Related Work 3 The Method 4 Experiments 5 Conclusions References Weakly Supervised Localization of the Abnormal Regions in Breast Cancer X-Ray Images Using Patches Classification 1 Introduction 2 Related Works 3 A Method of Weakly-Supervised Localization of the Cancer Regions in Breast Cancer X-Ray Images 3.1 Data Pre-processing 3.2 Generating Patches 3.3 Patch Augmentation 3.4 Patch Training 3.5 Heatmaps Calculation 4 Conclusions References Effects Evaluation of Data Augmentation Techniques on Common Seafood Types Classification Tasks 1 Introduction 2 Related Work 3 Method 3.1 Data Collection 3.2 Data Preprocessing and Augmentation Techniques 3.3 Split Data for the Evaluation 3.4 Building a Classification Model Using MobileNetV2 4 Experimental Results 4.1 Hyper-parameters Fine-Tuning 4.2 Effect of Data Augmentation Techniques on Seafood Types Classification 5 Conclusion References Image Caption Generator with a Combination Between Convolutional Neural Network and Long Short-Term Memory 1 Introduction 2 Related Work 3 Methodology 3.1 Preprocessing 3.2 Training with Deep Learning Approaches 4 Experimental Results 4.1 Dataset Description 4.2 Performance Comparison 4.3 Conclusion References Clothing Classification Using Shallow Convolutional Neural Networks 1 Introduction 2 Related Work 3 Dataset 4 Method 4.1 Overview 4.2 CNN-1 4.3 CNN-2 5 Experimental Results 6 Discussion 7 Conclusion References Similar Vietnamese Document Detection in Online Assignment Submission System 1 Introduction 2 Related Work 3 The Proposed Method 3.1 Data Preprocessing and Cosine Similarity Computation 3.2 The Semantic Similarity Between Sentences 3.3 Calculate the Similarity Between the Documents 4 Experimental Results 4.1 Data Description 4.2 Performance Evaluation 4.3 Similarity Between Sentences 4.4 Similarity Between Documents 4.5 The Application to Assignment Submission System 5 Conclusion References A Study of Causal Modeling with Time Delay for Frost Forecast Using Machine Learning from Data 1 Introduction 2 Causal Modeling with Time Delay 3 Input Variable Granulation 4 Implementation and Experiments 4.1 Datasets 4.2 Input and Output 4.3 Model Implementation 5 Experiments 5.1 Granulation for Air Temperature and Vapor Pressure 6 Conclusion and Future Work 6.1 Conclusion 6.2 Future Work References