1.Basics

Types of Machine Learning:

1. Supervised Learning: Models learn from labeled data, where each example has both input and output. The model makes predictions for new data based on the patterns it learned.

   • Examples: Regression, Classification.

2. Unsupervised Learning: Models identify patterns in unlabeled data, aiming to group similar data points or reduce data complexity without predefined categories.

   • Examples: Clustering, Dimensionality Reduction.

3. Reinforcement Learning: Models learn by interacting with an environment and receiving feedback (rewards or penalties) based on actions, adjusting their strategy to maximize total reward over time.

   • Example: Policy Optimization for training agents in game-playing.

Supervised Learning:

Supervised learning is when a model learns from examples that have both inputs and correct outputs. It uses these labeled examples to learn patterns so it can make accurate predictions on new, similar data.

Given data with students' IQ scores, CGPAs, and placement status (Yes/No), the model learns patterns between IQ, CGPA, and placement. Later, it can predict placement status for new students based on their IQ and CGPA.

Regression and classification

! 600

• Time Series Prediction: Predicts values over time based on past data (e.g., weather forecasting).

Types of Unsupervised Learning:

Clustering: is an unsupervised learning method where the model groups similar data points together based on patterns, without needing labeled outputs.

Example:

If we only have students' IQ scores and CGPAs (no placement status), clustering can group students into clusters, such as "High IQ & High CGPA" or "Low IQ & Moderate CGPA." This helps identify natural groupings of students, like those who might perform similarly in placement.

! 500

• Dimensionality Reduction: Reduces the number of features in data while retaining important information (e.g., PCA for image compression).

• Anomaly Detection: Identifies rare items that don’t fit common patterns (e.g., fraud detection).

Types of Reinforcement Learning:

• Value-Based Learning: Learns a value function to predict the best action in each state (e.g., Q-learning).
• Policy-Based Learning: Directly learns a policy that maps states to actions (e.g., Policy Gradient methods).
• Model-Based Learning: Builds a model of the environment to plan and simulate future actions (e.g., Dyna-Q in robotics).