ESDS Knowledge Base


What are the different types of machine learning?

The machine learning algorithms are usually classified as either- supervised or unsupervised.

Supervised Machine Learning Algorithms:

These can be applied to the learning carried out in the past to new data by the use of labelled examples for predicting the events in the future. It starts from an analysis of a known training dataset, and the learning algorithm produces an inferred function for making predictions regarding the output values. The system can provide targets for newer inputs after enough training. The learning algorithm can also compare the output with a correct and intended output and even find errors to modify the model.

Unsupervised Machine Learning Algorithms:

These are used when the information used for training is neither classified nor labelled. The unsupervised learning studies how the functions would infer a function for describing a hidden structure from unlabelled data. The system is unable to figure out the correct output, but is still able to explore the data and can draw inferences from the datasets used for describing the hidden structures from the unlabelled data.

Semi-supervised Machine Learning Algorithms:

In this type of learning method, the use of both- labelled and unlabeled data is used for training purposes. Usually, a small amount of labelled data and a large amount of unlabeled data is present. The systems using this method can considerably improve the accuracy in learning. This is chosen when the acquired data label requires the skilled as well as relevant resources to train or learn it. The acquisition of any unlabelled data doesn’t need any additional resources.

Reinforcement Machine Learning Algorithms:

It is a learning method that interacts with the environment by producing actions and discovers errors or rewards. The trial and error search and delayed reward are the essential characteristics of reinforced learning. This method allows machine and software agents to determine the ideal behaviour within a specific context to maximise the performance. Simple reward feedback is required in case of agents who are learning regarding the best action, known as the reinforcement signal.

Leave a Reply