Unit 6
1. What is the difference between Artificial Intelligence and Machine Learning?
2. What are the different types of machine learning techniques (supervised, unsupervised,
reinforcement)?
1. Supervised Learning:
Definition: Supervised learning is a type of machine learning where the model is
trained on a labeled dataset, which means the input data comes with corresponding
output labels.
Goal: The goal is for the model to learn the relationship between the input and output
data, so it can predict the output for new, unseen data.
Process: The algorithm learns from the training data by finding patterns and mapping
inputs to their correct outputs, refining its predictions over time.
Examples:
o Classification: Predicting categorical outcomes (e.g., email spam detection,
image recognition).
o Regression: Predicting continuous values (e.g., stock price prediction, house
price estimation).
Algorithms:
o Linear Regression
o Logistic Regression
o Decision Trees
o Random Forests
o Support Vector Machines (SVM)
o k-Nearest Neighbors (k-NN)
o Neural Networks
2. Unsupervised Learning:
Definition: Unsupervised learning is a type of machine learning where the model is
given input data without labeled outputs. The goal is to identify hidden patterns or
structures in the data.
Goal: The aim is to explore the underlying structure or distribution in the data and
extract meaningful insights without predefined labels.
Process: The algorithm tries to identify similarities or groupings in the data, or reduce
the dimensionality of the data.
Examples:
o Clustering: Grouping data points into clusters based on similarities (e.g.,
customer segmentation, grouping news articles).
o Dimensionality Reduction: Reducing the number of features in the data while
retaining important information (e.g., Principal Component Analysis - PCA).
Algorithms:
o k-Means Clustering
o Hierarchical Clustering
o DBSCAN (Density-Based Spatial Clustering of Applications with Noise)
o Principal Component Analysis (PCA)
o Independent Component Analysis (ICA)
o Gaussian Mixture Models (GMM)
3. Reinforcement Learning:
Definition: Reinforcement learning (RL) is a type of machine learning where an
agent learns to make decisions by performing actions in an environment to maximize
some notion of cumulative reward.
Goal: The agent aims to learn a policy that will maximize its long-term rewards, often
in an environment where actions have delayed consequences.
Process: The agent takes actions based on its current state, receives feedback in the
form of rewards or penalties, and adjusts its strategy over time to improve its
performance.
Examples:
o Gaming: Training an AI to play games like Chess, Go, or Dota 2.
o Robotics: Teaching robots to navigate and perform tasks autonomously.
o Self-Driving Cars: Enabling cars to make driving decisions in real-time to
maximize safety and efficiency.
Algorithms:
o Q-Learning
o Deep Q Networks (DQN)
o Policy Gradient Methods
o Monte Carlo Tree Search (MCTS)
o Actor-Critic Methods
Summary Table:
Type of
Definition Goal Examples Algorithms
Learning
Supervised Learning from To predict output Spam detection, Linear
Type of
Definition Goal Examples Algorithms
Learning
Regression,
labeled data for new data stock price
Decision Trees,
Learning (input-output based on learned prediction, and
Neural Networks,
pairs). patterns. image recognition.
SVM.
To identify Customer
Learning from K-Means
patterns, segmentation,
Unsupervised unlabeled data to Clustering, PCA,
groupings, or anomaly detection,
Learning find hidden DBSCAN,
reduce data and data
structures. GMM.
complexity. compression.
Learning by
To maximize Gaming (e.g., Q-Learning,
interacting with an
Reinforcement long-term reward AlphaGo), DQN, Policy
environment and
Learning through trial and Robotics, Self- Gradients, Actor-
receiving
error. driving cars. Critic.
feedback.
How do machine learning models learn from data?
✅ What It Does:
AI + Robotics manage:
Sorting
Picking
Packing
Scanning
💡 Example:
Flipkart and Amazon use AI bots to move goods around efficiently.
🎯 Benefits:
Faster order fulfillment
Lower human error
Reduced labor costs
🔐 7. AI in Supply Chain Visibility and Tracking
✅ What It Does:
Tracks product movement at every step
Alerts in case of delays, theft, or damage
Visual dashboards powered by AI for real-time insights
💡 Tools:
AI + Blockchain for secure tracking
Smart tags (RFID, QR codes)
🎯 Benefits:
Improves transparency
Builds trust with customers and partners
Helps with compliance and auditing
✅ Summary: Key Benefits of Using AI in Supply Chains
AI Use Case Business Impact
Demand Forecasting Reduces overstock and stockouts
Inventory Optimization Cuts holding costs and improves availability
Route Planning Lowers fuel and delivery time
Predictive Maintenance Prevents costly downtime
Supplier Risk Analysis Avoids risky or unreliable vendors
Warehouse Automation Speeds up operations, reduces labor costs.
Real-Time Tracking Increases control, trust, and visibility
🚀 Real-World Companies Using AI in Supply Chains:
Company AI Application
Amazon Robotics in warehouses, demand prediction, and delivery optimization
Flipkart AI-based route planning and inventory forecasting
Walmart ML-driven replenishment and stock management
DHL / FedEx AI for route planning, delivery time prediction, and package scanning
Tata Steel Predictive maintenance and smart logistics
✨ Final Thoughts:
Using AI in supply chains isn't just about cost-cutting—it's about making operations
intelligent, agile, and resilient. It helps businesses:
Predict disruptions before they happen
Respond to market demands faster
Deliver a better customer experience
7. What are some real-world examples of machine learning and improving business decision-making?
Vanshika! Here's a very detailed explanation of how Machine Learning (ML) is used in
real-world businesses to improve decision-making, with in-depth examples across
industries, how they work, and what benefits they bring. Let’s break it down step by step:
🌟 What Is Machine Learning in Business Decision-
Making?
Machine learning enables businesses to learn from data and make intelligent decisions
without being explicitly programmed. Instead of relying solely on human guesswork or fixed
rules, ML:
Finds hidden patterns in data
Predicts future outcomes
Automates complex decisions
This leads to faster, smarter, and more accurate business strategies.
🔍 Real-World Use Cases by Industry (In Detail)
🛒 1. Retail & E-commerce (Amazon, Flipkart, Myntra)
📦 Problem:
“How can we suggest the right product to the right customer and avoid running out of stock?”
🔍 Machine Learning Used For:
Product Recommendations
o Algorithms track what you view, click, and buy.
o ML models (like collaborative filtering or deep learning) suggest what you're
most likely to buy next.
Dynamic Pricing
o Adjusts prices in real-time based on demand, competition, time, and inventory.
Inventory Forecasting
o Predicts which products will be in demand next month or during festivals.
✅ Decision-Making Improved:
Which products to stock more or less
Personalized marketing to individual users
Pricing strategies during peak/off-peak seasons
💰 2. Banking & Finance (HDFC, SBI, Paytm)
📦 Problem:
“How can we predict who will repay a loan and who might commit fraud?”
🔍 Machine Learning Used For:
Credit Scoring Models
o ML evaluates credit history, salary, spending habits, etc., to score loan
applicants.
Fraud Detection
o Uses anomaly detection or classification algorithms to flag unusual
transactions.
Customer Segmentation
o Divides customers into high-risk, medium-risk, and low-risk groups for
personalized services.
✅ Decision-Making Improved:
Which customers are safe to offer loans or credit cards
What transactions to flag or block
How to design risk-based interest rates
🏥 3. Healthcare (Apollo, IBM Watson Health)
📦 Problem:
“How can we detect diseases earlier or avoid unnecessary hospitalizations?”
🔍 Machine Learning Used For:
Disease Prediction
o ML analyzes patient history, symptoms, and lab reports to predict diseases like
diabetes or cancer.
Medical Imaging
o Uses deep learning to detect tumors in X-rays or MRIs faster than humans.
Hospital Readmission Prediction
o Predicts if a patient is likely to come back after being discharged.
✅ Decision-Making Improved:
Early and more accurate diagnosis
Custom treatment plans for patients
Reducing hospital costs and improving care
🚚 4. Logistics & Supply Chain (DHL, Amazon Logistics, Zomato)
📦 Problem:
“How can we deliver faster and cheaper?”
🔍 Machine Learning Used For:
Route Optimization
o ML finds the most efficient delivery paths based on weather, traffic, and
delivery windows.
Demand Prediction
o Predicts where packages will need to go or where food will be ordered.
Fleet Management
o Uses predictive maintenance models to avoid vehicle breakdowns.
✅ Decision-Making Improved:
How to schedule deliveries for minimum cost and maximum speed
When and where to send delivery agents
When to maintain or replace delivery vehicles
🎯 5. Marketing & Advertising (Netflix, Instagram, Spotify)
📦 Problem:
“How do we attract the right audience and keep them engaged?”
🔍 Machine Learning Used For:
Customer Segmentation
o Groups users by age, location, purchase history, or interests.
Predictive Analytics
o Predicts which users are likely to click, buy, or churn.
Personalized Campaigns
o Automatically sends emails or push notifications tailored to each user.
✅ Decision-Making Improved:
What content or product to promote to whom
Which users are worth retaining and how to engage them
What time is best to send marketing messages
🏭 6. Manufacturing (Tata Steel, Bosch)
📦 Problem:
“How can we avoid production failures and improve efficiency?”
🔍 Machine Learning Used For:
Predictive Maintenance
o Monitors machine vibrations, temperature, and pressure to predict
breakdowns.
Defect Detection
o Uses image recognition to catch faulty products on the assembly line.
Process Optimization
o Learns from data to optimize energy use, material use, and time.
✅ Decision-Making Improved:
When to schedule maintenance to avoid shutdowns
Whether to reject or approve a manufactured item
How to lower the production cost per unit
7. Hospitality & Travel (OYO, MakeMyTrip, IRCTC)
📦 Problem:
“How do we get more bookings and avoid cancellations?”
🔍 Machine Learning Used For:
Dynamic Pricing
o Price rooms or tickets based on demand, day, and past booking trends.
Review Analysis
o Uses NLP (Natural Language Processing) to analyze reviews and improve
services.
Cancellation Prediction
o Predicts if a booking is likely to be cancelled and adjusts room availability.
✅ Decision-Making Improved:
What price to offer at any given time
Which properties to promote or demote
How to avoid overbooking or underbooking
🔁 Bonus: How ML Works in These Cases
1. Supervised Learning
Used when historical data has labels (e.g., fraud = yes/no)
Example: Predicting loan default, disease detection
2. Unsupervised Learning
Finds hidden patterns in unlabeled data
Example: Customer segmentation, anomaly detection
3. Reinforcement Learning
The model learns by trial and error to maximize reward
Example: Dynamic pricing, delivery route optimization
🧠 Final Takeaways:
Machine Learning Helps Businesses To: Resulting Benefits
Make sense of massive data. Smarter and faster decision-making
Predict future trends and customer behavior. Higher profits and better service
Automate repetitive and complex decisions Saves time, cost, and human effort
Improve customer satisfaction through
More loyalty and better retention
personalization.
Safer financial and operational
Reduce risk and fraud
environment
🚀 Want This As...
✅ A Slide Deck for presentations?
✅ An Infographic for notes?
✅ A Personalized Example for your jewelry business?
