Machine learning, often abbreviated as ML, is the study of algorithms that improve themselves by learning from large data sets. ML algorithms work by building models trained using an extensive data collection, known as training data. Then these models make predictions or decisions on new data, known as testing data.
ML algorithms are beneficial in applications like computer vision, where conventional algorithms cannot cope with the complexity.
Machine learning is associated with artificial intelligence and deep learning.
What are the approaches to machine learning?
- Supervised learning: In this learning method, the algorithm takes a large data set that contains inputs and their desired outputs. The training aims to use the training data to identify a function that correctly predicts the desired outcome for new data not in the training dataset.
- Unsupervised learning: In this learning method, the algorithm takes an extensive data set of inputs. The training aims to identify common patterns in the test data itself. Thus the algorithm can classify any new data based on the presence or absence of the patterns identified by the algorithm.
- Reinforcement learning: In this learning method, the algorithm takes a goal. Based on its actions, it gets either rewarded or punished by the environment. These actions and their consequences train the algorithm to predict the path towards the goal.