Machine Learning - Supervised Learning
The Supervised Learning Algorithm goal is to learn patterns in the data in order to build a general set of rules to map input to the class or an event.
We have two types of Supervised Learning algorithms:
1) The first one is Regression - these kind of algorithms are used to predict an output with continuos number, based on a given input dataset. For example Regression is used to predict financial stock exchanges.
2) The second one is Classification - the output has to be the probability of an event / class. The number of a class or an event can be two or more. These kind of algorithms learn the pattern in data to classify input to a discrete number of class outputs. For example we use Classification to take a dataset of images (for example to discriminate cats and dogs) and separate the ouput in two classes.
How to build Supervised Learning models
Building a supervised learning model in Machine Learning has three stages:
1) Training phase - we take an input dataset, and we want it to map with an output. The algorithm has to learn the patterns in order to map the input with the output. The output will be represented with a statistical equation, known as a model.
2) Testing - In this phase we evaluate the training of the dataset to predict the class or event.
3) Prediction - in this pahse we apply the trained model to our dataset, so we can figure it out if the prediction will be used or not.