Machine Learning Basics

Machine Learning Terminology

• Features are unique characteristics or details that help identify or describe something, such as a car's color, model, and year.
• A sample is a single piece of data used to study or learn from, like a photo, sound clip, or spreadsheet row.
• A feature vector is a list of feature values representing an object, such as a car's feature vector being ["Red", "Toyota", "2015"].
• Feature extraction is the process of simplifying or transforming data to make it easier to work with, like reducing an image's resolution.

Machine Learning Concepts

• Learning is the process where a machine learning model improves its predictions based on feedback.
• Intelligence is the ability to make correct decisions or predictions, such as diagnosing a disease based on symptoms.
• Artificial Intelligence is when a computer is programmed to make intelligent decisions, like diagnosing diseases.

Importance of Machine Learning

• Machine learning is useful in situations where it's hard for humans to explicitly program correct behavior or when the correct behavior might change over time.
• Examples include programming a car to drive or a spam filter to identify spam emails.

Machine Learning Process

• The learning system model involves a training set, predictions, and updates based on performance.
• Training and testing involve using a training set to teach a model and a testing set to evaluate its performance.

Types of Machine Learning

• Machine learning is a field of Artificial Intelligence that tries to make computers learn from data.
• Learning means improving at a task based on experience, such as a spam filter learning to identify spam emails.
• Types of learning include:
  • Rote learning: simple memorization
  • Learning from instruction: being told what to do
  • Learning by analogy: applying learned concepts to new contexts
  • Learning from observation and discovery (unsupervised learning): learning from data without explicit instructions
  • Learning from examples (inductive learning): learning from specific examples and generalizing to new situations

Inductive and Deductive Learning

• Inductive learning involves learning general rules from specific examples.
• Deductive learning involves using known facts and rules to learn new rules.