If this video is helpful to you, you can support this channel to grow much more by supporting on patreon : https://www.patreon.com/artofengineer Hello everyone, In this video, we will be discussing the types of algorithms in Machine Learning. There are many algorithms in Machine Learning and many more are yet being developed. But all of the Machine Learning algorithms can be classified into 3 main categories. And they are 1) Supervised Learning 2) Un Supervised Learning, and 3) Reinforcement Learning Supervised learning: Supervised Learning algorithms are used when the given data has both independent variable/s and target variable. And the task of the supervised learning algorithm is to find the relationship independent variable/s and the target variable. Let's try to understand this with a simple example. Let's assume we have some data about the hours studied by students and the marks scored by them in the exam. Here the hours studied is the independent variable and marks scored is the target variable. We are required to find what kind of relationship exists between the hours studied and marks scored. If we just plot the given data on the graph, we can see that the marks scored increases as the number of hours studied increases. Also we can see that there is linear relationship between the input and target variables. Thus the given data has a linear relationship which is of the form y = mx +c In Supervised learning, again there are two categories: 1) Algorithms for Regression type problems 2) Algorithms for Classification type problems When the target variables have continuous values, Regression type algorithms can be used. Example, the data of hours studied and marks scored. Here the marks scored is a continuous value like 30, 42.5, 53 etc., Thus, here regression type algorithms should be used. When the target variables have discrete or fixed set of values, we can use classification type algorithms. For example, if we have the data of hours studied and passed in exam or not, the target variable passed in exam or not has only 2 values pass or fail. Thus, we can use classification type algorithms here. Some of the popular supervised learning algorithms are linear regression, logistic regression, k nearest neighbours, support vector machines, decision trees, random forests and naïve bayes. Unsupervised Learning: Unsupervised Learning algorithms can be used when there are no target variables. In this case we are not trying to predict anything. We are just trying to find the pattern in the data if there are any. Let's take a simple example to understand this. We have the data which has two independent variables. One is Gender and other is Interested in either beauty products or gaming products. We don’t have any target variable in this case. If we can visualize this data, this is how it looks. We can observe some sort of pattern here. Many males are interested in gaming products. Many females are interested in beauty products. Very few females are interested in gaming products and very few males are interested in the beauty products. Basically when we feed this data to an unsupervised learning algorithm, the algorithm will simply find this pattern and output four clusters from this data and also given some random names to these clusters like cluster 1 , cluster 2 , cluster 3 and cluster 4. Once the pattern is ready, it can be very helpful. For example, let's assume we have a beauty product for men. We can easily target this product for cluster 4 instead of targeting on all the people. This will reduce the advertising costs significantly. Some of the popular unsupervised learning algorithms are KMeans Clustering, Heirarchical clustering and Single Linkage Clustering. Reinforcement learning: Reinforcement learning is all about letting an agent to interact in an environment. The agent will be rewarded for doing positive work and punished for doing negative work in the environment. Lets assume you are training your dog to do certain trick. Whenever your dog does the trick right, you will reward it by giving some treat. If the dog does the trick wrong, you will punish it by not giving any treat. The dog will understand this and tries to perform the trick correctly so as to maximize the positive reward.