Start With Machine Learning On Microsoft Azure


Machine Learning is used in everyone’s life every day. You probably have a question mark in your life of how certain predictable things happen, how corporations deal with it, or even how the government gives a predictive statement for every use case of our life in real time. Let's take as examples a Petrol price that gets hiked or gets cheaper; predictive use cases on the cost of gold; or a share market that ends up in giving a loan to a customer from the bank ranking towards their CIBIL score. Now, all these question marks come to an end with one solution named “Machine Learning”.

Here, in this article, I will be dealing with what Machine Learning on Microsoft Azure is and another series of articles following this will be based on the same Machine Learning on Microsoft Azure. We will be working with Machine Learning Studio using Microsoft Azure. Why do corporations rush to move to Machine Learning? We will be working on gathering the data, creating machine learning solutions, evaluateing the predictive solution from your machine learning sketch with the given data, and deploying it via Azure and maintaining the solution.

Why Azure Machine Learning?

Azure Machine Learning helps us to drag and drop the sketch where we design our solution with the Machine Learning Studio and this also helps us to work with code like Python, etc. It helps us in scaling and making the users access the predictive solutions that we have designed across the world.

What is Machine Learning?

As mentioned before, Machine Learning is used on many things in real life. Banks use machine learning with the help of the data given to predict the income of a particular client, debit statement records, credit statement records, and to make decisions on which loan the client is eligible for and what could be the money value that can be given to the client as a loan and what are the other loans that the client will not be eligible to access. Corporations like Flipkart use machine learning to learn which product you have searched a maximum number of times and help you in fetching the data which is related to your search history. Machine Learning simply works with the given data by the user and makes predictions and follows the instructions that have been given by the user who develops that particular Machine Learning solution.

Here, in Machine Learning,  we don’t use any traditional Control Logic, like if – case – while – until. Instead, we go with Machine Learning Logic. We gather the data that we need – sketch an algorithm which the machine learning can use – the algorithm analyzes the data – and create a model which is a solution to solve the problem based upon the data. So, the whole concept of Machine Learning depends upon the input data that has been given.

Techniques in Machine Learning:

Machine Learning goes with two techniques in learning data - Supervised and Unsupervised.
Supervised Machine Learning –  Supervised Machine Learning works with the data that has been provided on each field. Let's take an example. We set an input data with prices of cars such as 4 seaters and 6 seaters, with the cost in a particular city. Now, the Supervised Machine Learning algorithm will train the model with what would be the cost of the car with the given input data of a particular brand, using the training logic that has been given.


Unsupervised Machine Learning – Let's take the same scenario as the above one. In Unsupervised Machine Learning Techniques, this UML will be working with the given input, but it makes sense in such a model with attributes- like what sort of car the user is looking for in terms of engine model, engine power, cylinders, etc., and it provides results with the help of the same shades which get reflected with the help of UML algorithm. This also helps with the voice of the user, defined within the given input data.


I conclude this part of the article with the beginning of what Machine Learning is in detail. Follow my next article to know about Machine Learning in Azure, Machine Learning Studio, etc.