Machine learning techniques: supervised vs. unsupervised

Supervised learning

  • This is a technique of machine learning where labeled datasets are used to train an algorithm to determine the likelihood of a future event happening based on past occurrences of such an event.

Some real-world applications

  • To predict whether a customer will click on an ad or not.

  • Weather forecast: The machine will be fed conditions that led to rainfall and the conditions that did not. The algorithm then learns and can predict the probability of it raining.

Unsupervised learning

  • Unsupervised learning is a technique of machine learning where hidden patterns are uncovered from unlabelled datasets.

The clustering algorithm is commonly used to determine these hidden patterns by learning the structure of the unlabeled data, extracting useful information from it, and grouping similar data.

Anomaly Detection is another technique for unsupervised learning. It is used to find outliers in a collection of data points. It is also used to determine correlated data points in order to make business-based decisions.

Some real-world applications

  • Movie recommendation: The unlabelled data is fed to the algorithm. The algorithm then tries to group the data based on given parameters (movie genre, producer, duration). This can then be used to target different types of users based on their interests.

  • Detecting fraudulent transactions: Anomaly Detection is applicable.

You can say hi to me on Twitter!