Azure and Azure ML
In this article, we are going to learn about Azure ML and ML Studio. As we know Azure is Microsoft Cloud computing service. And Machine learning supported by Azure is called Azure ML. It’s a complete automated framework to build, teach, train and deploy as a web service and have visual development environment to make it easy for data scientists.
Benefits of having Azure ML as a cloud solution,
- It’s providing Azure ML Studio which it uses to create model and deploy instantly.
- Visual user interface, drag and drop feature, real time data visualization.
- All projects, experiments are stored in the cloud. You can access it from anywhere.
- Almost all input data types are supported by Azure ML as data source.
- Extend model using R and Python or use trained model as module.
Supported Input data types
- Hive
- HTTP
- MySQL
- SQL Server
- PostgreSQL
- Power Query
- SharePoint
- Azure DB
- Web API
- Local Files
- Teradata
There are many other data sources; you can check it on the Microsoft portal.
Azure ML Studio
Here I will give an overview of Azure ML Studio. You need to sign up in Azure portal and select Machine Learning Studio to launch it. It opens in the browser and looks like the below image.
It’s workbench software which has predefined protocols to follow while building and training a model. As per the image, the visual workspace enables developers to quickly create models and visualize data with just some clicks.
It has 6 high level navigations menus and those are Projects, Experiments, Web Service, Dataset, Trained Model and Settings.
Projects
It lists all projects and models created by users. Project contains combinations of all module experiments and datasets.
Experiments
It allows developers to build, test and iterate multiple times on either its new model or existing model. You can copy models and do many experiments and get accurate predictive results.
Web Service
Tested and trained models are deployed as web services as public APIs to use outside of the Azure environment. It predict results based on input parameters. It returns value based on trained deployed model data.
Dataset
Dataset contains uploaded datasets in Azure ML studio. It lists uploaded datasets and you can also pick from Microsoft sample datasets, which can be utilized for your experiments. You can use big New + buttons to add data files from your local computer.
Trained Model
Save your trained models and experiment for future uses.
Settings
Settings tab allows us to view and edit workspace and regenerate authorization token.
It also allows users to collaborate by inviting other users to clone or edit your experiments.
Task Bar
New + Button
New button is used in various new actions related to every process. New using dataset is used for uploading new data from your local computer. It also is used to upload preprocess modules developed in other language like Python or R for Modules. From New Experiments, you can pick sample experiments or create new empty experiments for projects. Other tool bar buttons are changed based on selected Navigation menu and items and processes.
Hope this gives you a high level overview of Azure ML studio. Thank for reading. Happy programming.