- Sheikh Aman
Machine Learning types | Types of learning in machine learning | THIRD one is used for robots.
Updated: Aug 10
what is machine learning?
Machine learning is a technology that deals with programmed systems. It gives results by automatically learning and experience. It works with less or no human interruption. Its main purpose is to construct an algorithm that will predict the output from all previous inputs.
The inputs into an algorithm is a training data or set and output is knowledge or you can say expertise, that form another algorithm which performs tasks.
Inputs can be numerical, audio, textual, multimedia, or visual.
Outputs can be floating number, category representation, classification, recognition like an apple from the food basket.
Machine Learning and its types
There are basically three types of learning.
Supervised Learning
Unsupervised Learning
Reinforcement Learning
Supervised Learning
It is commonly used for real-world projects. Examples- face verification, speech recognition,
product and movie recommendations and may more.
types of supervised learning
Regression
Classification
Regression- It trains and predicts a continuous type of values such as predicting carbon dioxide emission from cars etc.
Classification- It finds proper or you can say, appropriate class by analyzing sentiment like positive and negative, genders like a male, female or transgender, etc.
In this Learning, data for learning comes with a proper description or desired output and our objective is to form a general rule which maps all given inputs to that desired output.
Learning data is called labeled data.
Supervised Learning algorithms such as
Examples of Supervised Learning
spam and non-spam email classification
voice recognition
webpage labeling
Unsupervised Learning
It is used for anomalies detection, fraud detection or defective equipment detection, or grouping customers with similar behavior, etc. Opposite os supervised learning no labeled data.
when data contains very few indications and there is no description and labels. Then it depends on coder or algorithm to find a proper structure of undefined data, to determine the hidden patterns. This learning is called unlabeled data.
Example of Unsupervised Learning- There is a number of data points and we have to classify them but we don't know the criteria of classification. So, here an unsupervised algorithm suggests an optimum way for classification.
Unsupervised Learning algorithms
and so on...
Reinforcement learning
This learning is very interesting. In this learning, Learning data always gives feedback, which helps the system to adjust with the dynamic condition and achieve the required results.
After that system judges its performance based on the responses i.e feedback and then reacts according to that.
Example of Reinforcement Learning
self-driving cars
AlphaGo the chess master algorithm
Some important link only for you
A Great video By EDUREKA. Watch this for more information and clarification.
