Machine Learning Basics

Machine learning lets systems learn from data instead of being explicitly programmed. Three learning types dominate:

1. Supervised Learning

Learns from labeled examples: input → known output. Applications: spam detection, image classification, price forecasting. Algorithms: decision trees, random forest, neural networks.

2. Unsupervised Learning

Finds structures without labels: clustering (k-Means), anomaly detection, dimensionality reduction (PCA).

3. Reinforcement Learning

An agent learns through reward and punishment. Applications: games (AlphaGo), robotics, autonomous driving.

Key Concepts

Training vs inference, overfitting, train/test split, metrics like accuracy and F1. For deep networks see LLM Basics.