Types Of Machine Learning
We know that machine learning is the computer program which provides the ability to the machine to learn from the provided data, improves performance with experience, and making decisions or prediction. we have seen this in the previous post. Before going to the types of machine learning, I will give a brief on the process of programming or program solving and the machine learning model development process.
Traditional programming VS Machine learning
- In traditional programming, we develop a program in a certain language that contains a set of rules and data. After successful compilation and running, we will get a desired output or result. Hence the traditional programming means developing certain rules by the programmer and come up with desired output by providing the data.
- In machine learning, we have to feed the data which contains the input as well as output data that is the desired result. By using this data the model will train and come up with certain rules or parameters. So, by using that rules, we can predict output for the data which is unknown to the model.
This is the basic difference between traditional programming and machine learning. The figure given below will give a pictorial representation of the above concept.
Now we are familiar with the difference between programming and machine learning and the steps in machine learning development. Now, let’s see the types of machine learning.
Types of machine learning
By using the type of data provided to the model and the outcome of the model, they are divided into three major classes. The steps to build a machine learning model are the same in all types of machine learning. I will give a brief on each one by one.
- Supervised machine learning
- Unsupervised machine learning
- Reinforcement machine learning.
Supervised Machine Learning
In supervised machine learning, the data which we have to feed to the model should be labeled data. Labeled data means the data is in the pair of input and output data. So, we will feed the data with input and output. The model will train based on the training data. Then we have to test the model and check the accuracy of the model. For that, we will use data that is unknown to the model which is called test data. The following image illustrates this concept much better.
Based on the output data there are again two categories in supervised machine learning. If the output data is discrete then the model is considered in regression, and the data is categorical then it will be considered in classification. Let’s see what is regression and classification.
- Regression: Regression means the output data is in discrete form. There are some examples of regression such as home price prediction, semester marks prediction based on study hours, etc. In regression, we have to find the link which will best fit the given data. Again there are two small categories of regression is linear regression and non-linear regression. If the data is in a linear format and then the best fit line will be a straight line. If the line is not a straight one then it is called non-linear regression. I am not talking about any technical thing now. I will make another post as we go on. Linear regression, support vector machine are some algorithms for regression.
- Classification: If the output data is categorical, and there are some classes in the output data then it is better to go for classification. We have to feed the data and come up with the prediction of the class of the data. There are many classification algorithms such as decision tree algorithm, support vector machine, logistic regression, k nearest neighbors, etc. I will explain all these algorithms in upcoming posts as we go in deep.
Now, we have the basic idea about supervised machine learning, regression, and classification.
Unsupervised Machine Learning
In unsupervised machine learning, we have to feed the data which is unlabelled. Unlabelled data means the data without output. We will feed these data to the model and the model will generate the categories by analyzing the properties. A good example of unsupervised machine learning is clustering.
- Clustering: In clustering, the output of the clustering is the categories or classes.
Reinforcement Machine Learning
Reinforcement learning is a feedback-based learning method, in which a learning agent gets a reward for each right action and gets a penalty for each wrong action. The data provided to the model is also non-labeled data. In simple words, we can say that the output depends on the state of the current input and the next input depends on the output of the previous input. In Reinforcement learning decision is dependent, So we give labels to sequences of dependent decisions. You will get more clarification from the image given below.
The problems of reinforcement learning are such as chess playing game, robot wants to reach at diamonds without touching the obstacles. In these games, models have to act accordingly by analyzing the previous step.
Conclusion
In this post, I have given a brief on the difference between traditional programming and machine learning. Also, given information on the types of machine learning in brief.
Thank you.
Originally published at https://omkarraut.substack.com.
