Types of Machine Learning problems
Types of Machine Learning problems:
Supervised
Un-Supervised
Reinforcement Learning
Supervised Learning: Here, we want to make certain predictions for the future. Hence, we want the machine to learn the previous historical data and forccast for the future for instance, temperature for today. We provide both the input value and output (labels) of historical data such as climate details, humidity etc and along with that output like temperature. Hence, the model can find the derivatives between input and output and generate an equation. eg regression model.
Unsupervised Learning : USL is to simply find out different patterns in data and categorize something Or group something or segment something. Here, we only provide input values not output values. We will only provide the relative features of the class but will not label them. For instance, for identifying a dog, we give long tail, sharp teeth, sharp claws, makes a boow noise etc. But we will not give the label for it. We only expect the machine to make a segregation based on the underlying features. Therefore, Unsupervised Learning does not make any predictions for the future but only makes segregation.
Reinforcement Learning: it is known as reward-based learning. For instance, we have a robot that is learning how to walk. We train the robot that walk straight and if you strike anything on your way like wall or table etc, then turn left or right and move forward or come back. Each time he strikes somewhere, we would point out his mistake and tell him where he is going wrong. Reward him in such a way that if we does not strike anywhere, he provide a rewarded system such that the robot does not commit the same mistake. So this is what Reinforced Learning is all about.
Check out our master program in Data Science and ASP.NET- Complete Beginner to Advanced course and boost your confidence and knowledge.
URL: www.edujournal.com