MACHINE LEARNING

INDEX

Sr. No Title Page No.

1. Introduction 2

2. Learning Categories 3

3. Basic Distinction 4

4. How does ML Work 5

5. Restrictions of machine learning 6

6. Application Of ML 6

7. History of ML 7

8. ML at Present 9

9. Conclusion 10

10. Reference 11

1
MACHINE LEARNING

Introduction
Machine learning is a branch of artificial intelligence (AI) and computer science
that concentrates on the usage of data and algorithms to emulate the way that humans
learn, slowly enhancing its precision.

IBM has a rich history with machine learning. One of its own, Arthur Samuel, is
credited for coining the term, “machine learning” with his research (PDF, 481 KB)
(link resides outside IBM) about the game of checkers. Robert Nealey, the self-
proclaimed checkers master, played the game on an IBM 7094 computer in 1962, and
he lost to the computer. Corresponded to what can be done today, this feat seems
trivial, but it’s believed a main milestone in the area of artificial intelligence.

Over the last couple of decades, technological advances in storage and

processing power have allowed some creative products based on machine
learning, such as Netflix’s recommendation engine and self-driving cars.

learning is a significant component of the growing field of data science. Through

the usage of statistical methods, algorithms are trained to make categories or forecasts
and to reveal key insights in data mining projects. These insights subsequently drive
decision-making within applications and businesses, ideally influencing key growth
metrics. As big data continues to grow and grow, the market need for data scientists
will increase. They will be needed to help determine the most relevant business
questions and the data to answer them.

2
MACHINE LEARNING

❖ Machine learning can be categorized into two broad learning tasks:

1. Supervised ML
2. Unsupervised ML
There are numerous other algorithms.

1. Supervised learning:

An algorithm utilizes training data and feedback from humans to understand the
relationship between given inputs to a given output. For instance, a practitioner can
utilize marketing costs and weather forecasts as input data to forecast the sales of
cans. You can utilize supervised learning when the output data is known. The
algorithm will forecast new data.

There are two types of supervised learning:

1. Classification task
2. Regression task

Classification

Suppose you want to forecast the gender of a customer for a commercial. You
will start collecting data on height, weight, job, salary, purchasing basket, etc. from
your customer database. You know the gender of each of your customers, it can only be
male or female. The objective of the classifier will be to allocate a chance of being a
male or a female (i.e., the label) based on the information (i.e., features you have
gathered). When the model learned how to identify males or females, you can utilize
new data to make a prediction. For instance, you just got new information from an
anonymous customer, and you want to know if it is a male or female. If the classifier
forecasts male = 70%, it means the algorithm is sure at 70% that this customer is a
male, and 30% it is a female.
The label can be for two or more classes. The above Machine understanding
example has only two classes, but if a classifier requires to forecast an object, it has
dozens of classes (e.g., glass, table, shoes, etc. each object represents a class).

Regression

When the output is a continuous value, the task is a regression. For instance, a
financial analyst may require to forecast the value of a stock based on a range of
features like equity, last stock performances, and macroeconomics index. The system
will be trained to evaluate the price of the stocks with the lower possible error.

3
MACHINE LEARNING

2. Unsupervised learning

In unsupervised learning, an algorithm examines input data without being given

an explicit output variable (e.g., explores customer demographic data to recognize
patterns).
You can use it when you do not know how to organize the data, and you want the
algorithm to find patterns and categorize the data for you.
Example: Training of students during exams. While preparing for the exams students
don’t really cram the subject but try to learn it with full understanding. Before the
examination, they provide their machine(brain) with a good quantity of high-quality
data (questions and answers from different books or teachers’ notes, or online video
lectures). Even, if they are training their brain with input as well as output i.e. what
type of strategy or logic do they have to solve various types of questions? Each time
they solve practice test papers and find the performance (accuracy /score) by
comparing answers with the answer key given, Slowly, the performance keeps on
growing, achieving more confidence with the adopted method. That’s how actual
models are built, train the machine with data (both inputs and outputs are given to the
model), and when the time comes to test on data (with input only) and execute our
model scores by comparing its answer with the actual output which has not been
provided while training. Researchers are performing with assiduous efforts to
enhance algorithms, and methods so that these models perform even better.

Basic Distinction in ML and Traditional Programming?

• Traditional Programming: We provide in DATA (Input) + PROGRAM (logic), run

it on the machine, and obtain the output.
• Machine Learning: We provide in DATA(Input) + Output, run it on the machine
during training and the machine creates its own program(logic), which can be
evaluated while testing.

What does exactly learning mean for a computer A computer is said to be

learning from Experiences with regard to some class of Jobs if its performance
in a given job enhances the Experience.
A computer program is said to learn from experience E with respect to some class
of tasks T and performance measure P, if its performance at jobs in T, as measured
by P, enhances with experience E Example: playing checkers. E = the experience of
playing numerous games of checkers T = the task of playing checkers. P = the
possibility that the program will win the next game In general, any machine
learning problem can be appointed to one of two broad categories: Supervised
learning and Unsupervised learning.

4
MACHINE LEARNING

❖ How does ML work?

Machine learning is the brain where all the learning carries out. The way the
machine learns is identical to the human being. Humans learn from experience. The
more we know, the more efficiently we can forecast. By analogy, when we face an
unknown condition, the probability of success is lower than the known situation.
Machines are trained the same.
To create a precise forecast, the machine sees an example. When we give the
machine an identical example, it can figure out the outcome. However, like a human,
if it provides an earlier unseen example, the machine has complications predicting.
The core objective of machine learning is learning and inference. First of all, the
machine learns via the discovery of patterns. This discovery is created thanks to the
data. One essential part of the data scientist is to select carefully which data to
deliver to the machine. The list of attributes utilized to solve a problem is called a
feature vector. You can think of a feature vector as a subset of data that is utilized to
tackle a problem.
The machine utilizes some fancy algorithms to facilitate reality and transforms this
discovery into a model. Therefore, the learning stage is utilized to represent the data
and summarize it into a model.
• Collecting past data in any form appropriate for processing. The better the rate of
the data, the more suitable it will be for modeling
• Split the input data into training, cross-validation, and test sets. The ratio between
the respective sets must be 6:2:2
• Building models with suitable algorithms and methods on the training set.
• Testing our conceptualized model with data that was not provided to the model at
the time of training and assessing its performance utilizing metrics such as F1
score, accuracy, and recall.
• Linear Algebra
• Statistics and Probability
• Calculus
• Graph theory
• Programming Skills – Languages such as Python, R, MATLAB, C++, or Octave.

5
MACHINE LEARNING

❖ Restrictions of Machine Learning:

1. The prior challenge of machine learning is the shortage of data or the
diversity in the dataset.
2. A machine cannot learn if there is no data available. Similarly, a dataset
with a shortage of diversity gives the machine a hard time.
3. A machine requires to have heterogeneity to learn significant insight.
4. It is rare that an algorithm can remove information when there are no or
few deviations.
5. It is advised to have at least 20 observations per group to help the machine
learn. This constraint leads to poor evaluation and forecast.

➢ Application of Machine Learning:

Now in this Machine learning tutorial, let’s learn the applications of Machine
Learning:

➢ Augmentation:
Machine learning, helps humans with their day-to-day duties, personally or
commercially without having complete control of the output. Such machine learning
is utilized in various ways such as Virtual Assistants, Data analysis, and software
solutions. The preliminary user is to decrease mistakes due to human bias.

➢ Automation:
Machine learning works completely autonomously in any field without the
necessity for any human intervention. For example, robots execute the necessary
process steps in manufacturing plants.

➢ Finance Industry
Machine learning is expanding in popularity in the finance industry. Banks are
mostly using ML to find patterns inside the data but also to stop fraud.

➢ Government organization
The government makes use of ML to manage public security and utilities. Take
the example of China with its massive face recognition. The government utilizes
Artificial intelligence to prevent jaywalking.
➢ Healthcare industry
Healthcare was one of the first industries to use machine learning with image
detection.

6
MACHINE LEARNING

❖ History of Machine Learning

Before some years (about 40-50 years), machine knowledge was science fiction,
but today it is a part of our everyday life. Machine learning is making our day-to-day
life effortless from self-driving cars to Amazon virtual assistant "Alexa". However, the
idea of machine learning is so old and has a long history. Below some milestones are
given which have appeared in the history of machine learning:

➢ The earlier history of Machine Learning (Pre-1940):

o 1834: In 1834, Charles Babbage, the father of the computer, created a device
that could be programmed with punch cards. However, the machine was never
built, but all modern computers depend on its logical structure.
o 1936: In 1936, Alan Turing gave a theory that how a machine can decide and
perform a set of instructions.

➢ The generation of stored program computers:

o 1940: In 1940, the first manually managed computer, "ENIAC" was invented,
which was the first electronic general-purpose computer. After that stored
program computers such as EDSAC in 1949 and EDVAC in 1951 were
developed.
o 1943: In 1943, a human neural network was modeled with an electrical circuit.
In 1950, scientists started using their idea to work and analyzed how human
neurons might work.

➢ Computer machinery and intelligence:

o 1950: In 1950, Alan Turing posted a seminal paper, "Computer Machinery and
Intelligence," on the topic of artificial intelligence. In his paper, he questioned,
"Can machines think?"

7
MACHINE LEARNING

➢ Machine intelligence in Games:

o 1952: Arthur Samuel, who was the frontiersperson of machine learning,

developed a program that aided an IBM computer to play a checkers game. It
performed better more it played.
o 1959: In 1959, the term "Machine Learning" was first coined by Arthur Samuel.

➢ The first "AI" winter:

o The duration of 1974 to 1980 was a hard time for AI and ML researchers, and
this duration was called AI winter.
o During this duration, the failure of machine translation appeared, and people
reduced their interest in AI, which led to less funding by the government for the
research.

➢ Machine Learning from theory to reality

o 1959: In 1959, the first neural network was applied to a real-world problem to
extract echoes over phone lines using an adaptive filter.
o 1985: In 1985, Terry Sejnowski and Charles Rosenberg developed a neural
network NET talk, which was capable to train itself how to accurately
pronounce 20,000 words in one week.
o 1997: IBM's Deep blue intelligent computer won a chess game against the
chess expert Garry Kasparov, and it became the first computer that had beaten
a human chess professional.

8
MACHINE LEARNING

❖ Machine Learning at present:

Now machine learning has got an outstanding improvement in its research, and it
is present everywhere around us, such as in self-driving cars, Amazon Alexa, Catboats,
recommender systems, and numerous more. It contains Supervised, unsupervised, and
reinforcement learning with clustering, classification, decision tree, SVM algorithms,
etc.
Modern machine learning models can be utilized for creating various predictions,
including weather prediction, disease prediction, stock market analysis, etc.

9
MACHINE LEARNING

❖ Conclusion:

Machine learning is a significant component of the growing field of data science.

Through the usage of statistical methods, algorithms are trained to make categories or
forecasts and to reveal key insights in data mining projects. These insights
subsequently drive decision-making within applications and businesses, ideally
influencing key growth metrics. As big data continues to grow and grow, the market
need for data scientists will increase. They will be needed to help determine the most
relevant business questions and the data to answer them.

10
MACHINE LEARNING

❖ Reference :

• https://www.geeksforgeeks.org/machine-learning/
• https://www.ibm.com/think/topics/machine-learning
• https://mitsloan.mit.edu/ideas-made-to-matter/machine-learning-explained
• https://www.javatpoint.com/machine-learning

11