"Machine Learning is the field of study that gives computers the ability to learn without being explicitly programmed." — Arthur Samuel, 1959
A simple example is a spam filter. This can utilize Machine Learning because it is supplied with examples of spam emails and examples of regular emails, the program can then learn to identify a spam email on its own.
Some key terms to keep in mind:
- Training Set: The examples (data) the system uses to learn. Each of these examples is called a training instance.
- Accuracy: The way we measure our system's performance the percentage of predictions that are correct.
- Data Mining: Machine Learning techniques that analyze large amounts of data to discover patterns.
Machine Learning solves problems for which traditional programming might not find a good solution. ML-based solutions often outperform traditional ones because they can adapt to changes dynamically.
Types of Machine Learning Systems
There are three well-known types of systems classified by how they are trained:
- Supervised Learning: The system is trained with a dataset that includes the solutions (labels). For example, you have a training set of emails where each email is tagged with a label like "spam" or "not spam."
- Unsupervised Learning: The training data is unlabeled, the system tries to learn patterns without a guide.
- Reinforcement Learning: The system learns through trial and error, receiving rewards or penalties based on its actions.
