2025-09-16 22:30
Status: Tags: Machine Learning, Machine Learning Specialization
2 - Supervised and Unsupervised Learning
We discovered before that ML is a machine that learns. Then, you may ask, what it learns and how it learns.
To answer that, the way a machine learns is by being given a step-by-step instruction to achieve a goal; the so-called algorithm.
That is how, but what about the what? What does it learns? It is learning from the data that we give to them (the machine with an instruction).
Think about when we sit in our favorite class. Our teacher gave us a material to study, this material is the “data”. So we need to learn that material, and our teacher hopes that when the test comes, we can answer it correctly.
Okay, now we know how and what the machine learns.
If you remember, how our teachers teach us is somehow different from each other. There are types of teachers who give examples and analogies, and others only give the material without a quiz, etc.
That is also the case with how the machine learns. There are types of learning algorithms. The major and most used learning algorithms in real-world applications are Supervised Learning and Unsupervised Learning
Supervised Learning
Input x and output y pair, or given input x and the right answer y pair.
Regression
infinite possible output
Classification
Limited possible output
Unsupervised learning
Given input , we do not provide the right answer (y), and we might find some interesting data as an output.
Clustering (k-means)
Grouping the data by similar variables.
Anomaly Detection
Detect an unusual pattern in the data
Notes This is an active document, I’ll update this document when my knowledge about it is updated.