SUPERVISED LEARNING

Definition

Category of machine learning where algorithms learn to make predictions or classify data based on labeled training dataset.

Characteristics

Predictive Modeling


Labelled data

Clear Objectives

Human Intervention


Categories

CLASSIFICATION

REGRESSION

Binary Classification

Types

Definition

A method where the model tries to predict the correct label of a given input data.

Spam/ Not Spam emails

Matched / Not Matched Face Recognition

Churn/ Stay for a service

Positive/ Negative Sentiments

Multi-class Classification

Sports/ Technology/ Politics/ Entertainment in News articles

Positive/ Negative/ Neutral in feedback analysis

Defective/ Good/ Needs Repair in fault detection in manufacturing

Cat/ Dog/ Bird/ Elephant in image detection

Definition

A technique used to capture the relationships between independent and dependent variables, with the main purpose of predicting an outcome

Multiple Linear Regression

Logistic Regression

Types

Simple Linear Regression

Predicting if a credit card transaction is fraudulent or not fraudulent

Salary and Experience

Analyzing diseases based on variables like age, gender, weight, diet, exercise, and medication.

Algorithms

Support Vector Machines

Decision Tree

K - Nearest Neighbor

Simple to understand and implement.

Predictions are based on the majority class (classification) or average (regression) of the K nearest neighbors.

Sensitive to the choice of distance metric (e.g., Euclidean, Manhattan).

Classification and regression can be performed using KNN.

Applications of Supervised Learning

Recommendation System

Medical Diagnosis

Face Detection

Suitable for small to medium-sized datasets

Voice Recognition

Handwriting Recognition

Prone to overfitting with noisy data if not regularized properly

Objective of SVM is to find optimal hyperplane

Constructed by finding a hyperplane that maximizes the margin between different classes

Used for modeling decisions or rules in a tree-like structure.

Graphical representation for getting all the possible solutions to a problem/decision based on given conditions.

A decision tree simply asks a question, and based on the answer (Yes/No), it
further split the tree into subtrees.

Image-Based Diagnostics

Cancer Detection and Classification

Disease Detection

Archiving Handwritten Forms

Transcribing Medical Records

Handwriting-to-Text Apps

E-commerce Product Recommendations

Social Media Feeds

Movie and TV Show Recommendations

Music Recommendations

Voice Assistants

Voice Typing

Language Translation

Voice Biometrics

Face Detection in Cameras and Smartphones

Social Media Photo Tagging

Identity Verification

Agriculture

Crop Disease Detection

Crop Yield Prediction

Irrigation Optimization