My interest with Artificial Intelligence and specially with machine learning has always kept me searching for finding things that makes this field easy to grasp and understandable for masses. With the emergence of cloud computing and combining the Microsoft enterprise level applications tactfully a right combination has emerged in the form of Machine learning with Azure and that is taking the industry in the rightful direction which was needed by data scientist, BI professionals and DBAs. Until now data scientist has to identify the data set, then IT has to build an application in Python or R. This whole process could have taken months to complete a machine learning project.
With Azure ML(Machine Learning) the same application can be built in hours. Microsoft has given us Azure Machine Learning Studio which is a cloud based solution where data scientists can create their projects and experiments, publish and distribute their web services with so much ease that one should take a look and come back praising it. There is nothing to be installed on your laptop and don’t require anything from hardware aspect as everything is online and browser based. With all this you can do advanced learning and analytics.
Few years back when I was doing a project on Supervised learning with Artificial Neural Network and developing a multi-layered architecture with back-propagation technique it was tough to package and expose the solution in the form of a web service, where I can take the dataset in the form of input and create a specialized output after training the data and creating a prediction. But now with Azure ML most of the steps while creating a model has been given in the form of drag and drop. The user can setup properties of each step along with visualization. Thus after successfully creating an experiment we can easily setup a web service that can be beneficial for the community you want to share with. Even more you can publish your solution to Marketplace where you can monetize your efforts.
One of the incredible feature of Azure ML is the options of coding the application with R or Python or combining the both. Along with this there are builtin algorithms provided by Microsoft which consists of following high level categories:
- Recommender System
Each of these category has their sub-categories which is for fine-tuning the solution to the problem.
To check out Azure ML, visit:- https://studio.azureml.net
In the end, I am of the opinion that the direction Azure Machine Learning has taken will bring lot of predictive power to the businesses and communities thus churning more possibilities of solutions.