MACHINE LEARNING PROBLEM TYPES 'JAN 01-06-2025'

Machine Learning Problem Types

Machine learning problems are generally categorized based on the type of output the model is expected to produce. The two most common problem types are Regression and Classification, both of which fall under Supervised Learning.

Regression vs Classification
Aspect Regression Classification
Definition Predicts a continuous numerical value Predicts a categorical class or label
Output Type Continuous (numbers) Discrete (classes)
Examples of Output Price, temperature, salary Yes/No, Spam/Not Spam, Disease type
Common Algorithms Linear Regression, Polynomial Regression Logistic Regression, Decision Trees, SVM, KNN
Evaluation Metrics Mean Squared Error (MSE), RMSE, R² Accuracy, Precision, Recall, F1-score
Use Cases and Real-World Examples
Regression – Use Cases & Examples

House Price Prediction: Estimating the price of a house based on size, location, and amenities

Sales Forecasting: Predicting future sales revenue

Weather Prediction: Estimating temperature or rainfall

Stock Price Prediction: Forecasting stock values

Example:
A real estate company uses regression models to predict house prices based on historical market data.

Classification – Use Cases & Examples

Email Spam Detection: Classifying emails as spam or not spam

Medical Diagnosis: Identifying whether a patient has a disease

Credit Approval: Approving or rejecting loan applications

Image Recognition: Classifying images (cat vs dog)

Example:
An email service provider uses classification algorithms to filter spam emails from users’ inboxes.

Summary

Regression predicts numerical values

Classification predicts categories or labels

Both are essential machine learning problem types with wide real-world applications across industries