In this article, we will see what machine learning is and how it differs from traditional programming.
The first question is "What is Machine Learning?"
Machine Learning can be defined as building a model from given inputs to make data-driven predictions. It means that the system learns how to solve a problem from example data rather than you writing a specific logic.
In traditional programming, we analyze the problem and then write code; this code reads the data and uses the control logic in the code to determine the correct parts of the program to execute; which then produces the correct result. In this type of programming, we use if statements, case statements, and controlled loops.
In machine learning, however, we do not write the control logic that produces the result. Instead, we gather the data we need and we modify its format into a form which machine-learning algorithms can use. We then pass this data to the algorithm. The algorithm analyzes the data and creates a model, which implements the solution to solve the problem, based on the data.
In machine learning, the value of the data that we use drives the prediction logic and not the logic tests written by a developer, as was the case with traditional programming.
Types of Machine Learning
Machine Learning algorithms learn from data by utilizing one of two primary techniques – supervised or unsupervised.
![Machine Learning In Azure]()
In supervised machine learning, each row of data has fields containing feature values and the value that we want the algorithm to predict.
In unsupervised machine learning, there are clusters of like data. The algorithm analyzes the input data and identifies groups of data that share the same properties.
| Supervised Machine Learning | Unsupervised Machine Learning |
| Used for value prediction | Identify clusters of like data |
| Needs sample data containing value being predicted. | Data does not contain cluster membership |
| Trained model predicts value in new data | Model provides access to data by cluster |
Machine Learning Infrastructure Requirements
A very important part of any Machine Learning solution is Training Models. However, apart from these training models, we also need an infrastructure that gets and maintains the data we use to create the training model.
Therefore, if we have to build a machine learning solution, we need to have infrastructure which:-
- Allows to deploy the solution so that applications can easily use its capabilities
- Ensure that only authorized users or applications get access
- Allow us to easily maintain the solution
Microsoft Azure is an ideal platform to meet the above requirements for building a machine learning solution.
![Machine Learning In Azure]()
Summary
In this article, we got the introduction of Machine Learning and its types. In the next article, we will see how to use Microsoft Azure for Machine Learning.