n🧠 Understanding Human and Artificial Intelligence
What is Intelligence?
Human Intelligence: Our mental ability to make decisions, solve problems, and learn
new things.
Artificial Intelligence (AI): A machine's ability to make decisions, solve problems, and
learn, mimicking human-like capabilities.
Examples of Machine Learning (ML)
Object Classification: Identifying objects, such as apples vs. strawberries.
Anomaly Detection: Identifying unusual data points from live data streams.
Deep Learning (DL) Examples
Object Identification: Identifying complex objects using features like feathers, wings,
and beaks to identify a bird.
Digit Recognition: Recognizing digits from images.
📊 Data Fundamentals
What is Data?
Data is information in any form. For example, a table containing information about fruits.
Each row contains information about a different fruit, described by certain features.
Features and Labels
Features: Columns of a table, such as name, color, and size in a fruit dataset.
Labels: Special features that provide meaning to the data, depending on the problem we
are trying to solve.
Example: Fruit Data
Fruit Color Price
Apple Red $1.8
Orange Orange $2
Fruit Color Price
Banana Yellow $1
Grape Purple $3
In this example, if we are trying to predict the fruit based on its color, then color is the feature,
and fruit name is the label.
Data Labeling: The process of attaching meaning to data.
Types of Data
Labeled Data: Data with attached tags or labels (e.g., name, type, number).
Unlabeled Data: Raw data without any tags.
Training and Testing Datasets
Training Dataset: A collection of examples used to train the AI model. Like a teacher
using examples to teach a class.
Testing Dataset: Used to test the accuracy of the model. The test is performed without
labeled data, and then the results are verified with labels.
🤖 Types of AI Models
AI models can be classified as follows:
Rule-Based
Machine Learning (ML)
Deep Learning
Rule-Based AI Models
Advantages:
Easy to understand
Transparency
Predictable
Disadvantages:
Not flexible
Hard to maintain
Machine Learning
Supervised Learning
Unsupervised Learning
Reinforcement Learning
Artificial Neural Networks
Convolutional Neural Networks
🍎 Supervised Learning
Supervised learning involves learning from labeled input data to produce an output that
classifies the data.
Example: Currency Coin Prediction
Problem Statement: Build a model to predict the coin based on its weight.
1 Euro weighs 5 grams
1 Dirham weighs 7 grams
1 Dollar weighs 3 grams
1 Rupee weighs 4 grams
Feature: Weights
Label: Currency
The model learns from labeled input data and produces an output to classify the inputs.
Sub-categories of Supervised Learning
Classification Model
Regression Model
Classification Model
Classification is a process of finding a function that helps in dividing the dataset into classes
(discrete values) based on different parameters.
Multi-Class Classification
Binary Class Classification
Example: Predicting Team Win
Use a classification algorithm to estimate the probability of a data point belonging to either
Team A or Team B.
Example: Classifying Houses Based on Budget
High Budget House: > $500K
Medium Budget House: $100K - $500K
Low Budget House: < $100K
Real-Life Examples
Weather Prediction: Will the weather be hot or cold tomorrow? The model would be
trained on historical weather data, including temperature information labeled as "hot"
or "cold".
Location
Season
High and low temperatures
Humidity
Email Spam Detection: Classifying emails as spam or not spam.
Modern email classifiers identify if an email is spam and have evolved into
categories such as social, advertisement, and notifications.
Regression Algorithms
Regression algorithms predict a continuous value based on the input variables.
Continuous values include temperature, price, income, and age.
Purpose
The purpose of regression problems is to come up with a mapping function based on the input
and output variables to find a correlation between dependent and independent variables.
Example: House Price Prediction
Predict the price (dependent variable) based on parameters like the number of bedrooms,
carpet size, and garage area (independent variables).
Example: Weather Forecasting
A regression algorithm will be used to predict the weather. The model will learn based on past
data and apply that knowledge to predict the future weather.
Example: Used Car Price Prediction
Predict the selling price of a car with the help of parameters like:
Fuel type
Years of service
Number of previous owners
Kilometers driven
Transmission type (manual/automatic)
This type of model will be of type regression since it will predict an approximate price
(continuous value) of the car based on the training dataset.
Classification or Regression?
Predicting whether a customer is eligible for a bank loan or not? Classification
Predicting weather for the next 24 hours? Regression
🧑🏫 Unsupervised Learning
Unsupervised Learning is a type of learning without any guidance. The machine is responsible
for discovering patterns, similarities, and differences on its own based on the unlabeled dataset.
Example: Supermarket Customer Analysis
Assume that we have a customer database with records of their products bought over a period.
A marketing manager decides to send a grocery offer message to those customers who buy
groceries regularly.
Customer ID Total number of items bought Total number of grocery items bought Gro
Note that there were no customers labeled as grocery shoppers and non-grocery
shoppers.
The model could discover patterns on its own and could come up with these two
clusters/groups.
Key Aspects
In unsupervised learning, the model has to process information without any labels. It analyzes
and processes the data to identify hidden patterns and attributes and then uses that to classify
the inputs into categories based on similarities.
Supervised vs. Unsupervised Learning
Feature Supervised Learning Unsupervised Learning
Data Type Deals with labeled data Deals with unlabeled data
Useful in real-world problems-like Useful in finding unknown patterns within data-like mak
Usefulness predicting the prices of an item number of observations from an experimental device.
Computing
Power Simpler More complex
Supervised or Unsupervised?
Netflix recommendations based on someone's watch history: Unsupervised
Facebook identifies your friend in a picture from an album of tagged
photographs: Supervised
Analyze bank data for suspicious-looking transactions and flag the fraud transactions
(Note that suspicious transactions are not defined in this case): Unsupervised
🐾 Reinforcement Learning
Reinforcement learning is a type of learning in which a machine learns to perform a task
through a repeated trial-and-error method.
Key Terms
Agent: The learner or decision-maker.
Action: What the agent can do in its environment.
State: The current situation of the agent.
Reward: Positive feedback for a correct action.
Punishment/Penalty: Negative feedback for an incorrect action.
Environment: The surroundings in which the agent operates.
Analogy: A Dog Learning to Fetch
State: The current position of the dog in the park.
Agent: The dog.
Action: The action the dog has to take to fetch the stick and bring it back to the owner.
Summary: Family of ML Models
Supervised Learning Unsupervised Learning Reinforcement Le
Goal Determine relationships through training Discover new patterns Learn by rewardin
🧩 Sub-categories of Unsupervised Learning Model
Clustering Model
Association Model
What is Clustering?
Clustering is a process of dividing the data points into different groups or clusters based on their
similarity between them.
Difference Between Clustering and Classification
Featur
e Classification Clustering
Uses predefined classes in which Finds similarities between objects and places them in the same c
Process objects are assigned differentiates them from objects in other clusters.
Example: Music Preferences
Jim enjoys listening to music. Jim likes to listen to music having a slow tempo and soft intensity
whereas he dislikes songs with a fast tempo and high intensity.
We have grouped all the songs having slow tempo and soft intensity into 1 cluster that
he likes
While, songs with a fast tempo and high intensity into another cluster
Now if he listens to a new song X with a slow tempo and soft intensity. Could you predict
will he like the song X or not?
The clustering model will be able to identify clusters based on some similarities or patterns
which are not defined in the input. Similar techniques are used in OTT platforms like
Netflix/Spotify for recommendations.
What is Association?
Association Rule is an unsupervised learning method that is used to find interesting relationships
between variables from the database.
Based on the purchase pattern of customers A and B, can you predict what will any Customer X
who buys bread will most probably buy?
