There are three types of Machine Learning
Supervised Machine Learning
It is Task-Oriented learning where you provide various input and output samples. the machine then tries to learn from the given samples and figure out a relation (tries to map between the input and output). At a certain point, it has learned too much from those samples so that it can predict nearly the correct output of the new input.
There are two types of Supervised Learnings
1. Regression
It deals with predictions related to Numerical Data. Here you have to predict a numerical (continuous) value as an output.
example -
you are given the sum of marks of a student over the last 7 years, and you have to predict the sum for this year.
2. Classification
It deals with predictions related to Categorical Data. Here you have to predict a category to which the data belongs on the basis of input.
example - Sentiment analysis
You are given the facial properties of a person and you have to identify his sentiments(happy, sad, angry).
Unsupervised Machine Learning
In this learning, we just have input but no output by which we can apply supervised learning. here the data is processed for all the possible ways to group it in various types called labels(classes), and assigning a class to a member of this data is called labeling.
example - Buying habits of people
here various people have various needs and interests and its hard to group them. by applying unsupervised learning, we can identify new trends in buying habits.
Reinforcement Learning
It is a mixed version of learning where improvements on a previously learned model are continuously made to improve it.
example - an action game goes hard as you play it.
I am not sure but this can be an example of Reinforcement Learning. When I play shadow fight 3 I found that it gives suggestions like - your opponent is learning your moves, use different moves. so I think that it could be related to reinforcement learning.
Top comments (0)