Types of machine learning

Types of machine learning

1 mins read

Machine learning is sub-categorized into 3 types:

  1. Supervised learning
  2. Unsupervised learning
  3. Reinforcement learning

Supervised learning – Train Me!

Supervised learning as the name suggests the presence of a supervisor as a teacher. It is learning in which we teach or train the machine using data that is well labelled which means data is already tagged with the correct answer. And then, the machine is given a new set of data i.e. input to ensure that a supervised learning algorithm analyzes the learning data and creates a proper outcome from labelled data.

            Training the machine:

While training the machine, data is divided into the ratio of 80:20 i.e. 80% as training data and rest as the testing data. In training data, we feed input and output from 80% data. The machine learns from training data only. We have different machine learning algorithms to build our model. By learning, it means that the machine is going to build its own logic.

After the machine is ready then it is good to be tested. At the time of testing, the input is given from the remaining 20% data which the machine has never seen before, the machine will predict some output and we will compare it with actual output and compute the accuracy.

Types of Supervised learning:

  • Classification

In this supervised learning where output is having pre-defined labels or distinct values. It can either be binary or multi-class classification.

In Binary classification model predictions can either be 0 or 1 i.e. True or False.

In the Multi-Class classification, the model predicts more than 1 class.

  • Regression

In this supervised learning, the output is having continuous value.

            Supervised learning algorithms:

  • Linear regression
  • Nearest neighbour
  • Decision tree
  • Gaussian Naive Bayes
  • Random forest
  • Support vector machines(SVM)


Unsupervised learning – I am self-sufficient in learning!

Unsupervised learning is the training of machine using data that is neither classified nor labelled and allowing the algorithm to act on that data with no guidance. Here the role of the machine is grouping unsorted data according to similarities, differences, and patterns without any previous training of data.

Unlike supervised learning, no mentor is provided that means no training will be provided to the machine. Therefore the machine is restricted to search for the hidden framework in unlabeled data by our-self.

For example, if the machine is given an input image of cats and dogs

which it has never been seen before, thus the machine has no idea about the features of a cat or a dog so the machine is not able to categorize it in dogs and cats.

However, though it is able to categorize them according to differences, patterns, and similarities i.e., we are able to categorize the above image into 2 parts. The first could contain all pics having dogs in it and the second part might contain all pics having cats in it. Here you did not learn something previously, which means no training data or even examples.

Types of Unsupervised learning:

1)         Clustering

It is where you want to find the inherent grouping in the data, like grouping customers based on their purchasing behaviour.

2)         Association

It is where you want to discover rules in order to describe a large portion of the data, like the customer that buy X also tends to buy Y.


Reinforcement learning – Hit and Trial

It’s the ability of an agent to work with the environment and find the best outcome. It uses the concept of hit and trial method. The agent is penalized or rewarded with a point for a correct or a wrong answer, and based on the positive reward points received the model trains itself. And once trained it can predict the results of the new input data.