"Machine learning is a core, transformative way by which we're rethinking how we're doing everything." Sundar Pichai, CEO of Google
What is Supervised Machine Learning?
Supervised learning is a method in which models are trained using labeled data. It requires supervision to train the model.
For example: Consider, we have an image of different types of fruits. The task is to identify the fruits and classify them accordingly. So to identify the image in supervised learning, both the input data as well as output data will be given, which means the model will be trained by the shape, size, color, and taste of each fruit. Once the training is completed, the model will be tested by giving the new set of fruits. With the suitable algorithm, the model will identify the raw data and predict the output.
Supervised learning is of two types: Regression and Classification
- Regression: Estimate continuous values.
- Classification: Identify a unique class.
What is Unsupervised Machine Learning?
Unsupervised learning is another machine learning method in which patterns are identified from the unlabelled input data. The goal is to find the structure and patterns from the input data. They do not need any supervision. They find a pattern on its own from the given data.
For example: In unsupervised learning, we will not provide any supervision to the model. The input data is given to the model and allows the model to find the patterns from the data. With the suitable algorithm, the model will be trained itself and divide the raw data into different groups according to the features similar to them.
Unsupervised learning is of two types: Clustering and Association
- Clustering: It splits the dataset into groups based on their similarities.
- Association: It allows you to establish associations amongst data objects inside large databases.
What is Reinforcement Learning?
When we compare these three types of learning, reinforcement learning is a bit different than the other two. To find out what is the best outcome an agent can interact with the environment. The concept of the hit and trial method is followed. According to the answer, whether correct or wrong, the agent is rewarded or penalized with a point and the model is trained itself based on the positive reward points gained. And once trained it gets ready to predict the new data given to it.
