1.
The Need and Objective of Machine Learning (ML)
ML is a promising and flourishing field. The objective of machine learning is
to process archival data for organizations to make better decisions, such as to
design new products, improve business processes, and to develop
effective decision support systems. Machine learning enables top
management of an organization to extract knowledge from the data stored in
various archives of the business.
Historically, the full potential of the huge amount of data used by business
organizations was not utilized due to two reasons:
1. Data scattering: Data was spread across different archive systems,
preventing organizations from fully integrating these sources.
2. Lack of software awareness: There was a lack of awareness regarding
software tools that could unearth useful information from the data.
2. Factors Contributing to ML's Popularity
Machine learning has become so popular because of three primary reasons:
1. High Volume of Available Data: Big companies like Facebook, Twitter, and
YouTube generate vast amounts of data that grows rapidly. It is estimated that
this data approximately doubles every year.
2. Reduced Cost of Storage and Hardware: The hardware cost has dropped,
making it easier to capture, process, store, distribute, and transmit digital
information.
3. Availability of Complex Algorithms: With the popularity and adaptation of
machine learning, especially with the advent of deep learning, many complex
algorithms are available.
3. Key Concepts: The Knowledge Pyramid (Figure 1.1)
Before embarking on the machine learning journey, it is necessary to establish
the terms—data, information, knowledge, intelligence, and wisdom. The
Knowledge Pyramid illustrates the progression of these concepts.
Level Definition Characteristics/Notes
Consists of numbers or text that can be processed
by a computer. It is mostly available as raw facts
and symbols (base of the pyramid). Organizations
Data All facts.
are accumulating vast and growing amounts of data
from sources like flat files, databases, or data
warehouses.
Includes patterns, associations, or relationships
Informati
Processed data. among data. Example: analyzing sales data to find
on
the fast-selling product.
Example: historical patterns and future trends
Knowled Condensed obtained from analyzed sales data. Data is useless
ge information. unless knowledge is extracted, and knowledge is not
useful unless it is put into action.
The applied
Intelligen This is the actionable form of knowledge. Computer
knowledge for
ce systems have been successful up to this stage.
actions.
Represents the
maturity of Wisdom is the ultimate objective of the knowledge
Wisdom
mind (the apex pyramid, exhibited only by humans so far.
of the pyramid).
Introduction to Machine Learning
Chapter 1: Introduction to Machine Learning
The Field of Machine Learning (ML)
Machine Learning (ML) is described as a promising and flourishing field. A
relevant quote summarizing the goal of such technology is: "Computers are able
to see, hear and learn. Welcome to the future" (Dave Waters).
Primary Utility/Objective of ML:
Machine Learning enables the top management of an organization to extract
knowledge from data stored in various business archives. The objective is to
process archival data so organizations can take better decisions.
Decisions facilitated by ML are useful for organizations to:
• Design new products.
• Improve business processes.
• Develop decision support systems.
Learning Objectives (Key Topics for Examination)
The topics covered by this introductory material include:
• Exploring the basics of machine learning.
• Introducing the types of machine learning.
• Providing an overview of machine learning tasks.
• Stating the components of the machine learning algorithm.
• Exploring the machine learning process.
• Surveying some machine learning applications.
1.1 Need for Machine Learning
Business organizations utilize a huge amount of data for their daily activities.
Historically, the full potential of this data was not utilized due to two main
reasons:
1. Data scattering and integration difficulties: Data was scattered across
different archive systems, and organizations were unable to fully integrate these
sources.
2. Lack of awareness of software tools: There was a lack of awareness
regarding software tools that could help unearth useful information from the
data.
Business organizations are now adopting machine learning, the latest
technology, to overcome these limitations and utilize the data effectively.
Reasons for the Popularity of Machine Learning
Machine learning has become popular due to three key developments:
1. High Volume of Available Data: Big companies like Facebook, Twitter, and
YouTube generate vast amounts of data that grow at a phenomenal rate; it is
estimated that this data approximately gets doubled every year.
2. Reduced Cost of Storage and Hardware: The cost of storage and
hardware has reduced. This reduction makes it easier to capture, process, store,
distribute, and transmit digital information.
3. Availability of Complex Algorithms: The availability of complex algorithms,
especially with the advent of deep learning, contributes to machine learning's
popularity.
Due to its popularity and ready adoption by business organizations, ML has
become a dominant technology trend.
The Knowledge Pyramid (Data, Information, Knowledge, Intelligence, and
Wisdom)
Before delving into ML, it is essential to establish the terms Data, Information,
Knowledge, Intelligence, and Wisdom (often structured as the Knowledge
Pyramid, shown in Figure 1.1 in the source material).
The Knowledge Pyramid illustrates the transformation from raw facts to ultimate
understanding.
Level in Status in Computer
Term Definition/Description
Pyramid Systems
All facts are data. Data is mostly
available as raw facts and symbols. It
Base Data Basis for processing.
can be numbers or text processed by a
computer.
Processed data. This includes
Second Informati patterns, associations, or relationships Result of processing
Level on among data. Example: Analyzing sales data.
data to find the fast-selling product.
Condensed information. Example:
The historical patterns and future
Third Knowled Extracted
trends obtained from sales data.
Level ge patterns/trends.
Knowledge extraction is necessary, as
unextracted data is useless.
Computer systems
Fourth Intelligen Applied knowledge for actions. It is have been successful
Level ce an actionable form of knowledge. up to this stage
(Intelligence).
The ultimate objective of the
So far, exhibited
Apex Wisdom knowledge pyramid, representing the
only by humans.
maturity of mind.