- Supervised machine learning uses a dataset containing both features (x) and a target (y) to learn a function that maps inputs to outputs.
- Because the algorithm sees the correct answers, it learns how the other variables relate to the target.
- Regression predicts a numerical target; classification assigns each example to one of two or more classes.
- Choosing between two classes is binary classification; choosing among more than two is multiclass classification.
- In supervised learning the model gets inputs and outputs; in unsupervised learning it gets only inputs.
What is Supervised Machine Learning?
Supervised learning is the common approach when you have a dataset containing both features (x) and target (y) that you are trying to predict. You apply supervised machine learning algorithms to approximate a function (f) that best maps inputs (x) to an output variable (y).
Because the machine learning algorithm was provided with the correct answer to the problem, the algorithm is able to learn how the other variables relate to the value/answer. This allows you to discover the insights that predict future outcomes based on historical patterns.
What are Examples of Supervised Machine Learning?
Below are two examples of Supervised Machine Learning:
- Regression:the algorithm produces a numerical target for each example, for instance, how much revenue will be generated from a new marketing campaign.
- Classification:the algorithm labels each example by deciding between two or more different classes. Making decisions between two classes is called a binary classification, for instance, predicting if a person will default on a loan or not. When a decision has to be made between more than two classes it is known as a multiclass classification.
Why Is Supervised Machine Learning Important?
The importance of machine learning lies in its ability to turn data into actionable insights. It allows businesses to leverage data to better understand and avoid an unwanted outcome, or to increase the desired outcome of a target variable.
Supervised Machine Learning vs Other Technologies & Methodologies
Supervised machine learning vs unsupervised
With supervised learning, input data is provided to the model along with the output. For unsupervised learning, only input data is provided to the model.
Supervised machine learning vs deep learning
In classical supervised models, high-level abstraction of input features is not created. But in deep neural networks, abstractions of input features are formed internally.
Supervised machine learning advantages and disadvantages
An advantage of supervised learning is its ability to collect data or produce a data output from the previous experience. A disadvantage of the model is that decision boundary might be overstrained if your training set doesn’t have examples that you want to have in a class.
Supervised Machine Learning in H2O AI
H2O-3 includes supervised algorithms such as GLM, GBM, Distributed Random Forest, XGBoost, Deep Learning and Stacked Ensembles. H2O Driverless AI, our AutoML, also trains models such as LightGBM, RuleFit and TensorFlow. H2O-3 supervised learning docs.