Machine Learning:

UNIT-I
Machine Learning Introduction: Rote Learning, Inductive Learning, Reinforcement
Learning, Unsupervised Learning, Supervised Learning. Well posed learning problems,
Designing a Learning system, Perspective and Issues in Machine Learning. Concept learning
task, Concept learning as search, Find-S algorithm, Version space, Candidate Elimination
algorithm, Inductive Bias. Decision Tree Learning: Decision tree representation,
Appropriate problems for decision tree learning, Basic decision tree learning algorithm,
hypothesis space search in decision tree learning, Inductive bias in decision tree learning,
Issues in decision tree learning. [13
Hours]
Machine Learning:
Machine learning is a field of study in artificial intelligence concerned with the development
and study of statistical algorithms that can learn from data and generalize to unseen data, and
thus perform tasks without explicit instructions.
In other words, machine learning algorithms use data to learn patterns, relationships, and
insights, which they can then use to make decisions or predictions on new, unseen data. The
goal of machine learning is to enable computers to learn from experience and improve their
performance over time without human intervention.
Definition 1:
The major aim of ML is to allow the systems to learn by themselves through experience
without any kind of human intervention or assistance.
Defn 2:
Machine learning is a subset of AI, which uses algorithms that learn from data to make
predictions.
Rote Learning:
"Rote learning" refers to a learning technique where information is memorized through
repetition without necessarily understanding the underlying concepts or relationships. It
involves simply repeating information over and over again until it is committee
d to memory, without necessarily understanding its meaning or context.
Example:
1. a student might memorize the formula for calculating the area of a rectangle (length ×
width) without understanding why the formula works or how to adapt it to find the
area of other shapes.
2. Another example of rote learning could involve memorizing historical dates and
events without understanding the broader context or significance.
Inductive learning:
An technique of machine learning called inductive learning trains a model to generate
predictions based on examples or observations.
Inductive learning Is the type of learning techniques in machine learning (ML) typically
refers to the process of learning patterns, structures, or rules from examples in order to make
predictions or decisions about new data.
Example:
1.
2. Spam Detection:
In email spam detection, machine learning algorithms can analyze the entire content of an
email to determine whether it's spam or not.
2 .Disease diagnosis:
Disease diagnosis is another application of Machine Learning .
Inductive learning has the ability to analyse large dataset of medical information ,including
patient history, lab results, imaging data.
Inductive learning can find the patterns and trends by examining this datasets that could be
challenging for medical professionals to spot on their own

Reinforcement Learning:
Reinforcement learning is a method in machine learning where an agent learns to make
decisions by interacting with an environment. The agent receives feedback in the form of
rewards or penalties for its actions, aiming to maximize its total reward over time.
Example:
Consider a self-driving car learning to navigate traffic. When it follows traffic rules and
avoids accidents, it receives positive feedback. But if it breaks rules or causes accidents, it
gets negative feedback. Through this feedback, the car learns to drive safely and efficiently,
just like how we learn to drive by observing the consequences of our actions on the road.
Reinforcement learning helps the car improve its driving behaviour over time based on the
rewards or penalties it receives.
Unsupervised Learning:
Unsupervised learning is a type of machine learning in which models are trained using
unlabeled dataset and are allowed to act on that data without any supervision.
Example:
Imagine you have a box filled with differently shaped blocks—some square, some
circular, and some triangular—but they're all mixed up. Without any labels or
instructions, ML models sort them into groups based on their shapes using unsupervised
learning algorithms.
Supervised Learning:
Supervised learning is a type of machine learning .It learns the relationship between
input and output. The inputs are known as features or ‘X variables’ and output is
generally referred to as the target or ‘y variable’. The type of data that contains both
the features and the target (input and output)is known as labeled data. Here the
algorithm is trained on a labeled dataset. Then the algorithm learns the relationship
between the input data and the output data. Once trained, the algorithm can predict or
classify new, unseen data based on its learned(trained) knowledge.
Example:
Example of supervised learning is language translation. The user should provide a
dataset to computer with pairs of English sentences and their corresponding French
translations. The computer learns from these labeled examples, capturing the patterns
and relationships between the two languages. After training, the model can be used to
translate new English sentences into French accurately, based on what it learned from
the labeled data. This application is valuable for services like online language
translation tools.
Well posed training problems:
Well-posed learning problems are ones that have clear goals, enough data, suitable methods,
and ways to measure success. Examples include predicting prices, sorting emails, recognizing
numbers, diagnosing illnesses, and suggesting products.
“A computer program is said to learn from Experience E when given a task T, and some
performance measure P. If it performs on T with a performance measure P, then it upgrades
with experience E.
The three important components of a well-posed learning problem are,
1. Task
2. Performance Measure
3. Experience
Learning to play Checkers:
A checkers learning problem involves playing the game of checkers. The goal is to win as
many games as possible against opponents. To improve, the program plays games against
itself, gaining experience and learning from each game to become better at winning.
To simplify,
T -> Play the checkers game.
P -> Percentage of games won against the opponent.
E -> Playing practice games against itself.
2.Handwriting recognition (HWR) is a technology that converts a user’s handwritten letters
or words into a computer-readable format (e.g., Unicode text).
Its applications are numerous, it is used in reading postal addresses, bank forms, etc.
T–
> recognizing and classifying handwritten words from images.
P -> Percentage of correctly identified words.
E -> set of handwritten words with their classifications in a database
Designing a Learning system:
Designing a learning system in machine learning involves several important steps:
1. Identify Data Type: Figure out what kind of data you're working with, like numbers,
categories, text, or images. This helps choose the right algorithms and prepare the data.
2. Define Outcome: Decide what you want the system to do, like classifying things, making
predictions, or finding patterns. This guides the choice of algorithms and how you measure
success.
3. Consider Resources: Think about what you have to work with, like how much data you
have, how powerful your computer is, and how much time you have. This affects which
algorithms you can use and how much data you can use.
4. Choose Algorithm: Pick a machine learning algorithm that suits your task, such as decision
trees, support vector machines (SVMs), or neural networks.

5. Train the Model: Use your data to teach the algorithm how to do the task. Make sure to
check how well it's doing using measures like accuracy, precision, and recall.
6. Fine-Tune: Adjust the model's settings to make it work even better. This can involve testing
different options and seeing which ones give the best results.
7. Test**: Finally, see how well your model performs on new data that it hasn't seen before.
This helps ensure it's really learning and not just memorizing the training data.
By following these steps, you can create a learning system that does what you need it to do
and keeps getting better over time.

Perspective and issues in Machine Learning

Perspectives:
1. Advancements: Machine learning helps in making big discoveries in different areas like
healthcare and finance by analyzing large sets of data.
2. Automation: It makes tasks faster and easier by doing things automatically that used to be
done by hand.
3. Personalization: Machine learning can make things more tailored to you, like suggesting
movies you might like based on what you've watched before.
4. Decision Making: It helps in making decisions by looking at data and finding patterns that
people might miss.
5. Innovation: Machine learning encourages new ideas and inventions by giving us new ways
to understand and use data.
Issues:
1. **Data Quality**: If the data used isn't good, the predictions made by machine learning
might not be accurate.
2. **Bias and Fairness**: Sometimes, machine learning can make unfair decisions because of
biases in the data it's trained on.
3. **Interpretability**: Some machine learning models are hard to understand, so it's tough to
know why they make the decisions they do.
4. **Privacy Concerns**: Using personal data for machine learning raises worries about
privacy and keeping information safe.
5. **Ethical Considerations**: We need to think carefully about how we use machine
learning and make sure we're doing it in a way that's fair and responsible for everyone.

Concept Learning:
Concept learning is a fundamental task in machine learning that involves automatically
discovering and generalizing patterns or concepts from data.
A machine is taught the different examples of a concept, and by learning these examples, it
will “discover” the general rule(s) that apply to that concept. Concept learning thus involves
learning a function (which is a rule) from a set of training examples.