Home » Blog » Semi-Supervised Learning | Machine Learning Beginners

# Semi-Supervised Learning | Machine Learning Beginners

Hello Learners,

In this article I am going to describe what semi-supervised learning is?

As a formal definition

Semi-supervised learning is an approach to machine learning that combines labeled data with large amount of unlabeled data during training.

Semi-supervised learning is falls between supervised learning and unsupervised learning. A semi-supervised learning problems starts with a series of labeled data points as well as some data points for which labels are not known.

The main objective of semi-supervised learning is to classify some of the unlabeled data using labeled information set.

We can further divide Semi-supervised learning into two section

• Learning from positive and unlabeled data
• Learning from labeled and unlabeled data

As an example of positive and unlabeled data, consider your web browser that contains many bookmarked web-pages. There are thousands of web-pages available on the internet. But you only choose few of theme which are usable to you and bookmarked them.

Suppose you have bookmarked 50 web-pages, and on the internet suppose there are 1000 web-pages. So there are 1000 – 50 = 950 web-pages for you which are unlabeled.

Now as an example of labeled and unlabeled data, consider you are using any movie streaming application. Now you have watched few of them and based on that you are giving rating for those movies.

So suppose there are 1000 movies in that application, you give positive review for 40 movies and negative reviews for 20 movies. So we can distinguish movies as two part.

One is rated movies and another is unrated movie. Rated movies also divided into how many positive reviews and how many negative reviews. So we are having 40(positive review) + 20(negative review) = 60 (rated movies). And We are having 1000(total movie) – 60(rated movies) = 940(unrated movies).

## Real World Example Of Semi-Supervised Learning

### Speech Analysis

This is classic example of semi-supervised learning. Labeling audio files is very intensive task that requires lot of human resources.

### Web content classification

Organizing the knowledge available in billions of web pages will advance segment of AI. Unfortunately, that task typically requires human intervention to classify the content.