In machine learning, there are different learning methods that are used when working with algorithm or models.
In this article, I will be writing on supervised learning and unsupervised learning. Some of the things I will be explaining are:
- The meaning of supervised and unsupervised learning
- Their pros and cons
- And their applications.
Supervised learning is a machine learning technique that deals with well labelled data.
It works on a ground truth, that is, having an understanding or knowledge of what an algorithm or model should do from the given data.
Just like having mentors or supervisors assigned to a project to serve as a guide on what the final project should look like, that's how supervised learning works, the labelled data and ground truth serves as a guide to what the developed model or algorithm should do.
Pros and Cons
In supervised learning, there is a set of reference points used in developing the model, there is already an understanding of what the model or algorithm should do and how it should do it from the data.
It is very easy to measure the accuracy of developed model and there is barely any room for mistakes or errors as it can be easily corrected.
But then, one disadvantage in using supervised learning is the availability of labelled data as most times it is very expensive and difficult.
Using this learning method, a model can be developed without knowing the natural structure or other features of the given date as only the labelled part is focused on.
Application of Supervised Learning
Supervised learning can be applied in classification and regression problems. Classification has to do with mapping input to an output, while regression works with continuous output. Supervised learning can be applied in:
- Database marketing
- Handwriting recognition
- Information extraction
- Object recognition in computer vision
- Pattern recognition
- Speech recognition
In this learning technique, no well labelled data or guide is provided for the model. It is one of the learning methods that is often unpredictable, as it develops a model or algorithm with no understanding of what it does.
It analysis and evaluates the natural structure and features present in the dataset.
Pros and Cons
Just as explained above, in using unsupervised learning method, it allows for complex task to be performed on a model as information on the features can be discovered.
Getting of data for unsupervised learning is very easy as compared to supervised data.
Unsupervised learning takes place in real time as you work with the model or algorithm.
It is very difficult to measure the accuracy of model to know of it is doing the correct thing or not as information was provided.
One application of unsupervised learning is in the field of density estimation in statistics, though it encompasses many other areas that involves summarizing and explaining data features.
Hope you enjoyed reading this, do like this article or leave a comment. Thank you