Coefficient of Variation and Machine Learning Applications

K. Hima Bindu, Raghava Morusupalli, Nilanjan Dey, C. Raghavendra Rao

• Publisher: CRC Press
• Year: 2019
• ISBN:
  9781000752625
• Edition: 1
• Topic: Computer Science
• Subtopic: Computer Science General
• Index: Computer Science

Coefficient of Variation (CV) is a unit free index indicating the consistency of the data associated with a real-world process and is simple to mold into computational paradigms. This book provides necessary exposure of computational strategies, properties of CV and extracting the metadata leading to efficient knowledge representation. It also compiles representational and classification strategies based on the CV through illustrative explanations. The potential nature of CV in the context of contemporary Machine Learning strategies and the Big Data paradigms is demonstrated through selected applications. Overall, this book explains statistical parameters and knowledge representation models.

CHAPTER 1

Introduction to Coefficient of Variation

1.1 INTRODUCTION

The growth in communication, Internet, and computing technologies is persistently posing problems to the researchers. The data getting collected are becoming unmanageable, and developments in this direction led to the paradigm of research known as “Data Science” or “Big Data.” The experience of automation for past two decades is quite encouraging and given a hope, as Machine Learning methods with computational intelligence systems are producing promising results. Hence, we are going to experience higher level of comfort, integrity, quality of life, and so on.

The sophisticated systems getting built with the help of these developments resulted in new life style, namely, e-governess, smart cities, healthcare, paperless transactions, mobility-free transactions, and so on. The smartness in performance (operations) of these systems is due to consideration of all variations in the design of the systems such that the recommended (adopted) decisions will be apt for the given context of the target state and environment.

It is time to relook at what the scientific methodologies can contribute to this futuristic systems architecture. The decision-making processes in general are associated with matching patterns that are already registered and then adopt to more or less the similar decision to current situation. The matching process is highly time consuming. To speed up this matching process, several methods are being designed and constructed. Most of these tools are based on data analysis.

Expert systems (decision support system and knowledge support system) provided tools for decision-making by encapsulating rules and built appropriate inference engines more in medical domain (Mycin [1] is an example). The tractability limitations of building knowledge and inference engines came in the way of growth in Artificial Intelligence and Expert Systems. Technological growth in databases, Data Mining, and pattern recognition tools gave a n...