What is Unsupervised Learning?

Unsupervised learning is a type of machine learning to deal with data without giving the right answers. To find out the hidden labels in the dataset, we use unsupervised learning.

Let’s say we have a dataset, but there is no labels on it.

However, when observing the dataset, we can find out that there are two clusters in the figure. Some data points are very close to each other and some of them are not. Therefore, we can label these data by how close they are. Now we have a dataset with two labels on it.

As a result, the dataset is given by some new labels based on what approaches we deal with the dataset. A well-known approach is to use k-means clustering. However, the choice of an approach is highly regard to what your data looks like.

Unsupervised learning is quite interesting and very helpful. Why is that? First, unsupervised learning could tell something that we don’t know. Second, we know that supervised learning needs a lot of data with true labels, but it is hard to have these data. With unsupervised learning, it can help us to label the dataset and throw into the computer to train it! Moreover, dealing with a dataset with both labeled and unlabeled data is called “Semi-supervised Learning”.

Related posts:

There is war across the stream, I’ve read. And corruption over the wall.. “There is War Across the Stream” is published by Jeffrey Steen.

When I decided to start my personal training business in 2015, I thought my Certificate III and IV in Fitness was all I needed to get started. I’d come from a successful career in welfare. One where…

1. Hyphenate your phrasal adjectives. A small-business incentive is different from a small business incentive. A limited-liability clause is different from a limited liability clause. When two or…