Soft Computing Techniques in Engineering, Health, Mathematical and Social Sciences

Pradip Debnath, S. A. Mohiuddine, Pradip Debnath, S. A. Mohiuddine

• Publisher: CRC Press
• Year: 2021
• ISBN:
  9781000409840
• Edition: 1
• Topic: Computer Science
• Subtopic: Computer Engineering
• Index: Computer Science

Soft computing techniques are no longer limited to the arena of computer science. The discipline has an exponentially growing demand in other branches of science and engineering and even into health and social science. This book contains theory and applications of soft computing in engineering, health, and social and applied sciences. Different soft computing techniques such as artificial neural networks, fuzzy systems, evolutionary algorithms and hybrid systems are discussed. It also contains important chapters in machine learning and clustering. This book presents a survey of the existing knowledge and also the current state of art development through original new contributions from the researchers. This book may be used as a one-stop reference book for a broad range of readers worldwide interested in soft computing. In each chapter, the preliminaries have been presented first and then the advanced discussion takes place. Learners and researchers from a wide variety of backgrounds will find several useful tools and techniques to develop their soft computing skills. This book is meant for graduate students, faculty and researchers willing to expand their knowledge in any branch of soft computing. The readers of this book will require minimum prerequisites of undergraduate studies in computation and mathematics.

1 Revisiting the Machine Learning Algorithms and Applications in Engineering and Computer Science

Satyabrata Nath, Purnendu Das, and Pradip Debnath

Assam University Silchar

Contents

1.1 Introduction

1.2 Different Learning Approaches Used by Machine Learning

1.3 Algorithms Used by Different Learning Approaches

1.3.1 Supervised Learning Algorithms

1.3.2 Unsupervised Learning Algorithms

1.3.3 Semi-supervised Learning Algorithms

1.3.4 Reinforcement Learning Algorithms

1.4 Analysis of the Learning Algorithms

1.5 Related Works in the Field of Engineering and Computer Science

1.6 Discussion

1.7 Conclusion

References

1.1 Introduction

Artificial Intelligence (AI) is an intelligence mechanism for machines which enables them to imitate human intelligence and mimic their behavior to a certain extent or better (Charniak 1985, Genesereth and Nilsson 2012). AI was developed keeping in mind to solve complex real-world problems through an approximation of human decision-making capabilities and perform tasks in ever more humane ways (Russell and Norvig 2002).

Machine learning is a subset of Artificial Intelligence that enables the system to learn from past experiences and adapt itself to improve without programming explicitly for each task (Mohri et al. 2018, Alpaydin 2020). This adaptability enables machine learning algorithms to perform operations like prediction or classification based on the available data as well as making them feasible and cost-effective than manual programming (Domingos 2012). Machine learning became more popular with the rise of the internet and the accumulation of abundant data in the form of digital information, which made the engineers realize that instead of teaching the computer/machines to perform every task, it would be more efficient if they connect the system to the internet where all the data is accessible, then it would be possible for the system to learn by itself (Jordan and Mitchell 2015). Due to this automation of systems, machine learning is widely used in various fields consisting of data mining, robotics, image and speech recognition, medical diagnosis, natural language processing, sentiment analysis, classification and prediction, fraud detection and many more.

There are generally three stages for conducting machine learning (Mitchell 1997) – The first stage is called Training Phase where the model is trained with a training dataset in which the inputs are mapped with expected outputs. This phase prepares the machine learning model for classification and prediction tasks. Stage two is the validation and testing phase where we test the model with some test samples and measure its performance that how well it has been trained. Measures like errors, accuracy, time and space complexities, etc. are observed. The third and final stage is the application stage where the model is introduced to the real world to extract necessary information and solve real-life problems.

If we look back into the past of machine learning origins, Alan Turing in the year 1950 developed the “Turing Test” to examine if a computer has human-like intelligence (Turing 1950). In the 1950s, Arthur Samuel created a computer program for playing checkers and made the term “machine learning” popular (Samuel 1959). Rosenblat invented “Perceptron” in 1957, an electronic device which worked on the principles of biological learning mechanism and simulated human thought process (Rosenblatt 1957). In 1967, the algorithm “Nearest Neighbor” was developed, enabling computers to start using very simple pattern recognition. This could be used to map a path for traveling salesmen who began at a random city but ensuring that they visit all cities within a short tour (Cover and Hart 1967). Kunihiko Fukushima introduced “Neocognitron” in 1979 which was a multilayered ANN and was used for the identification of handwritten characters and other tasks of pattern recognition and laid the groundwork for convolutional neural networks (Fukushima 1980). In 1982, John Hopfield popularized Hopfield networks, which could act as a content addressable network for recurring neural network applications (Hopfield 1982). In 1986, the backpropagation method was popularized for ANNs (Rumelhart et al. 1986), and in 1989, Christopher Watkins developed Q-learning, which dramatically increases the practicality and effectiveness of reinforcement learning (Watkins 1989). In 1995, two popular methods were introduced in the field of machine learning. A paper explaining random forest decisions was published by Tin Kam Ho (Ho 1995), whereas Corinna Cortes and Vladimir Vapnik have presented their research on Support Vector Machines (Cortes and Vapnik 1995). After that, many methods and algorithms were developed in this arena and are still advancing with the provision of accessing large data from everywhere, thereby efficiently utilizing machine learning...

Table of contents

• Cover
• Half Title
• Series Page
• Title Page
• Copyright Page
• Table of Contents
• Preface
• Editors
• Contributors
• Chapter 1 Revisiting the Machine Learning Algorithms and Applications in Engineering and Computer Science
• Chapter 2 Detection and Prevention of Cancer through Artificial Intelligence and Machine Learning
• Chapter 3 A Review on Soft Computing Techniques in Nanomagnetism and Its Impact on Biomedical Applications
• Chapter 4 A Mediative Fuzzy Logic-Based Approach to the Goal Programming Problem: In the Context of Multi-Objective Solid Transportation Problem
• Chapter 5 H-U-R Stability Results of Mixed-Type Additive-Quadratic Functional Equation in Fuzzy β-Normed Spaces by Two Different Approaches
• Chapter 6 Tauberian Theorems for Intuitionistic Fuzzy Normed Spaces
• Chapter 7 Picture Fuzzy Soft Matrices
• Chapter 8 Cubic n-Inner Product Space
• Chapter 9 Convergence Methods for Double Sequences and Applications in Neutrosophic Normed Spaces
• Chapter 10 Intuitionistic Fuzzy Generalized Lucas Ideal Convergent Sequence Spaces Associated with Orlicz Function
• Chapter 11 Soft Computing Techniques in Social Sciences: The Recent Developments
• Chapter 12 An Approach Based on Fuzzy Logic for Analysis on Product Development in Open Innovation Context
• Index