Machine Learning is a subfield of artificial intelligence.
Scientists have been trying to create intelligent machines. The machine was good at one and only one thing couldn’t do anything else besides the task it was programmed to do. Enter Machine Learning.
Related course: Python Machine Learning Course
In the year 1997 a computer called Deep Blue beat the world champion of chess playing chess.
Intelligent? No, that was all this computer could do: It couldn’t do anything else.
So if you have a new problem, the computer wouldn’t be able to solve it. That means creating a new program with new logic and rules.
Instead of relying on hard coded rules, you can use algorithms that learn from examples and experience. This is called machine learning.
Machine learning algorithms can solve many probems like:
- spam classification: if an e-mail is spam or not.
- document classification: what type of document is it.
- face detection: is face or not?.
- price prediction: what is the expected price?.
and many more.
There are different types of machine learning algorithms, they can be divided into:
- supervised learning algorithms
- unsupervised learning algorithms
A type of supervised learning problem is classification. Suppose an algorithm must decide if an email is spam or not. You can write many logical rules, but that’s troubling to maintain and difficult to change.
What you can do instead is create a program that uses labeled examples of outputs that should be produced for input. The program uses traning data. After the algorithm is trained with the training data, you can make predictions.
Given a new example email, the algorithm can predict/classify if a new email is spam or not spam. Because training data is used, we call it a supervised learning algorithm.