A4.1.1 Types of Machine Learning and Their Applications in the Real World Notes

What is Machine Learning?

1. Machine learning is a type of artificial intelligence where computers learn from data and improve their performance over time without being explicitly programmed.
2. Algorithms detect patterns, make decisions, and predict outcomes based on previous examples.

Supervised Learning

1. Supervised learning involves training a model on a labeled dataset, where each input is paired with a corresponding output.
2. Types:
   1. Classification: Predicts discrete categories (e.g., spam or not spam).
   2. Regression: Predicts continuous values (e.g., house prices).

Unsupervised Learning

1. Unsupervised learning deals with unlabeled data, aiming to find hidden patterns or structures.
2. Types:
   1. Clustering: Groups similar data points (e.g., customer segmentation).
   2. Association rule mining (finding item relationships)
   3. Dimensionality Reduction: Reduces data complexity while preserving important features (e.g., PCA).

Reinforcement Learning

1. Reinforcement learning involves an agent learning to make decisions by interacting with an environment to maximize cumulative reward.
2. Key Concepts:
   1. Exploration vs. Exploitation: Balancing trying new actions and using known strategies.
   2. Policy: A strategy that defines the agent's actions.

Deep Learning

1. Deep learning is a subset of machine learning that uses artificial neural networks with multiple layers to automatically extract and learn hierarchical features from large, complex data sets.
2. Architectures:
   1. Convolutional Neural Networks (CNNs): Ideal for image processing.
   2. Recurrent Neural Networks (RNNs): Designed for sequential data like text.
   3. Transformers – modern Natural Language Processing (NLP) and vision models

Transfer Learning

1. Transfer learning involves reusing a pre-trained model on a new, related task, with the aim to reduce the time and data needed for training a new model.
2. How it works:
   1. Use a pre-trained model (e.g. on a general dataset)
   2. Fine-tune it on a smaller, specific dataset
3. Benefits:
   1. Reduces training time.
   2. Improves performance with limited data.

Real-World Applications of Machine Learning

Market Basket Analysis

1. Description: Identifies associations between products in transaction data.
2. Algorithm: Apriori (Unsupervised Learning).