Are you wondering how to expand your business, analyse stakes, get rid of liabilities and marginalize profits? Supervised learning is the calculated solution to all your problems!!! You can rule the stock market by having this skill at your side.

Supervised learning is teaching the model with labelled data and later testing it. It means the current activity is being overseen or supervised to make sure that the results come out to be accurate and correct. Machine learns under guidance of the training data. The training data comprises of the input as well as the output to set an example of how the predicted output should look like. The data which has input as well as output is known as labelled data. Supervised learning uses labelled data. A real life example would be children under supervision of parents or teachers.

Supervised learning solves classification(predicting label,class or category) and regression problems(predicting a continuous variable).Supervised learning has a very well defined and explicit training phase. Hence, it predicts the outcome directly very efficiently . It maps labelled input to known output.It has a well defined feedback mechanism as the data is predefined.

How to detect credit card fraud ? How to classify the spam mails in your inbox ? How your needs match up to the recommendation list? Well unsupervised learning plays a part in all of this!!!

Unsupervised learning means the activity is performed independently without supervision. The unlabeled input data is provided. The model tends to derive patterns in the given data and learn from it. For example, you and me observe patterns in the current world and derive our own solutions to different challenges that we face. Unsupervised learning aims to derive patterns and get useful insights from this data like trends and associations. It explores and plays with data until it discovers the output.

Unsupervised learning solves clustering (target audience grouping),anomaly detection(and abnormal activity tracking) and association(finding patterns and co-occurrences in data) problems.It has no feedback mechanism.

Have you heard about the self driving car that has created buzz in the market ? Moreover, have you played the alpha go game and wondered how all of it is really possible ? Well then you should definitely have a look at reinforced learning.

Reinforcement learning is basically forced learning based on a hit and trial method . We experience then evaluate our actions and hence, learn from them. In other words in reinforced learning, an agent interacts with the unknown environment via actions and the stage change or transition results in rewards or errors or penalties. It works with no predefined data. As a result, it starts from scratch by collecting data in the environment . It learns from experience. Feedback is received in terms of rewards and punishments. Like getting new life or losing one in a game as a reward or punishment.

To conclude we can say that Machine Learning can be classified as supervised, non-supervised or reinforced which have been discussed keeping in mind their similarities and differences. Will be further editing this blog post to include more technical aspects of these classifications.