This document provides an overview of machine learning including:
- The four main types of learning are supervised, unsupervised, semi-supervised, and reinforcement learning. Supervised learning uses labeled training data, unsupervised learning uses unlabeled data to find hidden patterns.
- Classification, regression, and clustering are common machine learning tasks and outputs. Classification and regression are supervised tasks that predict labels or numeric values, while clustering is unsupervised and groups unlabeled data.
- Decision trees are a popular supervised learning algorithm that can be used for both classification and regression. They use entropy, information gain, and other metrics to split data into branches.
- The key difference between artificial intelligence and machine learning is that AI is the broader field of
A Secure and Reliable Document Management System is Essential.docx
Machine Learning Fundamentals Explained
1. Machine Learning
Created by Ashwin Shiv
Roll No:181210013
CSE 3rdYear,5th Semester
Colloquium/ Industrial Lecture/ Seminar(CSP 311)
1
2. Introduction
Machine Learning is the field of study that gives computers the capability to learn without being explicitly programmed.
General type of Learning System
• “Signal” or “Feedback” available to a learning system
• “Output” desired from a machine learned system
2
3. “Signal” or “Feedback”
Learning depends on Input/ Output data and label/ un-label data and small/Large
data set
• Supervised learning
• Unsupervised learning
• Semi-supervised learning
• Reinforcement learning
3
4. Supervised learning
The computer is presented with example inputs and their desired outputs. Fully
labelled input and output.
Example
Image Classification: You train with images/labels.Then in the future you give a new
image expecting that the computer will recognize the new object.
Market Prediction/Regression: You train the computer with historical market data
and ask the computer to predict the new price in the future.
4
5. Advantages and Disadvantages of Supervised
Learning
Advantages of Supervised Learning:
• Machine can predict data on the basis of experiences.
• Used in solving real life problems such as fraud detection ,spam filtering etc.
• Knowledge about classes and objects is sufficient
Disadvantages of Supervised Learning:
• If test data is different from training dataset, then it cannot predict the correct output.
• Not suitable for handling complex tasks.
5
6. Unsupervised learning
No labels are given to the learning algorithm, leaving it on its own to find structure in its
input.
Example
• Clustering:You ask the computer to separate similar data into clusters, this is essential in
research and science..
• Generative Models: After a model captures the probability distribution of your input data,
it will be able to generate more data.This can be very useful to make your classifier more
robust.
6
7. Advantages and Disadvantages of
Unsupervised Learning
Advantages of Unsupervised Learning:
• Helps to make our machine adapt to new type of data.
• Is preferable as it is easy to get unlabeled data in comparison to labeled data.
Disadvantages of Unsupervised Learning:
• More difficult than supervised learning as it does not have corresponding output.
• Result of Unsupervised Learning Algorithm might be less accurate as input data is
unlabeled and these algorithms do not any exact output in advance.
7
8. Semi-supervised learning
Problems where you have a large amount of input data and only some of the
data is labelled.
Example
• Image Labeling : Only few images are label and majority are unlabelled.
Used by Facebook and Google photos
8
9. Reinforcement learning
When you present the algorithm with examples that lack labels, as in
unsupervised learning. However, you can accompany an example with
positive or negative feedback according to the solution the algorithm
proposes comes under the category of Reinforcement learning
Example
• Driving a vehicle
• Machine playing a game against an opponent
9
10. “Output” desired system
Learning tasks arises when one considers the desired output of a machine-
learned system
• Classification
• Regression
• Clustering
10
11. Classification
Supervised Learning Problem in which when inputs are divided into two or more
classes, and the learner must produce a model that assigns unseen inputs to one or
more (multi-label classification) of these classes.
Example:
• Like in emails,whenever we receive an email,spam filtering takes place which in
turns recognizes whether that mail is spam or not spam.Spam and not spam are
defined as label inputs to detect whether that email is spam or not spam.
11
12. Regression
Supervised Learning problem in which it checks if there is a relationship between the
input variable and the output variable
Example:
• Predicting stock prices using historical data.
• Prediction of rain using temperature and other factors
• Determining Market trends
• Prediction of road accidents due to rash driving.
12
13. Clustering
An Unsupervised learning problem in which objects are grouped into clusters such
that objects with most similarities remain into a group.These clusters found
something common between data objects and are categorized by our machine.
Example:
• It is used by the Amazon in its recommendation system to provide the
recommendations as per the past search of products.
• Netflix also uses this technique to recommend the movies and web-series to its
users as per the watch history.
13
14. Applications of Machine Learning
• Image Recognition:Used to identify objects, persons, places, digital images etc.
Eg:Automatic friend tagging suggestion
• Traffic Prediction: Google Maps predicts traffic in two ways:
a) RealTime location of the vehicle form Google Map app and sensors
b) Average time has taken on past days at the same time.
• Spam and Not Spam Filtering on emails
• Self Driving Cars
• Product Recommendations
• Speech Recognitions
14
15. DecisionTree Classification Algorithm
• Supervised learning technique to solve decision problems.
• Used for both classification and Regression problems, but mostly it is
preferred for solving Classification problems.
• It is a graphical representation for getting all the possible solutions to a
problem/decision based on given conditions.
15
17. Advantages and Disadvantages of Decision
Tree
Advantages of DecisionTree:
• Helps to think about all the possible outcomes for a problem
• Less requirement of data cleaning compared to other algorithms
• Useful for solving decision-related problems.
Disadvantages of DecisionTree:
• Some decision tree might contain a lot of layers which will become very difficult to read.
• If more layers are defined, time complexity may increase which leads to poor performance.
17
18. Formulae to be Used for making a Decision
tree
• Entropy of Class: It is a metric to measure the impurity in a given attribute
• Information Gain: Denoted by I(Pi,Ni). Almost same as Entropy of Class
• Entropy of an Attribute:
• Gain=Entropy of Class-Entropy of Attribute
18
19. Difference between Artificial Intelligence and
Machine Learning
19
Artificial Intelligence(AI) Machine Learning(ML)
AI is the study of how to make
computers just like humans.
ML is a subset of AI which is defined as
field of study that gives computers the
capability to learn without being
explicitly programmed.
Includes learning, reasoning, and self-
correction.
Includes learning and self-correction
when introduced with new data.
Goal of AI is to make a computer
system that can think like humans
Goal of ML is to make our machine
learn which can provide accurate
output.
AI is concerned about maximizing the
chances of success.
ML is concerned about accuracy and
patterns.
AI has a very wide range of scope. ML has a limited scope.
AI will generally go for finding optimal
solution.
ML will go for solution depending on
the input given.