Continued from page 1.
Founded in 2011, BigML is a machine learning platform used primarily for predictive modeling. The BigML platform features anomaly detection, cluster analysis, SunBurst visualization for decision trees, text analysis, and more. The BigML API allows applications to access predictive models and other BigML resources. Using the API, applications can perform CRUD operations on BigML resources using standard HTTP methods.
BigML provides a nicely designed developer site that features well-organized and comprehensive API documentation, code samples, client libraries, a quick start page, and other developer tools. In February 2014, BigML reached a major milestone: 1 million predictive models created with the BigML platform.
Founded in 2013, PredictionIO is an open source machine learning server that makes it possible to quickly build predictive engines. PredictionIO features a variety of almost-complete engine templates that can be customized for use cases such as recommendation systems, sentiment analysis, document classification, search results ranking, and product ranking.
PredictionIO features an Event Server that can collect and store arbitrary events. Applications can send events to the server via API, and application events can be retrieved or deleted via API. PredictionIO provides a well-organized and comprehensive documentation site that features SDKs, developer guides, demo tutorials, and more. The latest version of PredictionIO (0.9 Series) was released in March and includes several major improvements, such as new engine templates, evaluation metrics, and hyperparameter tuning support.
Microsoft Azure Machine Learning
API Documentation URL: https://azure.microsoft.com/en-us/services/machine-learning/api/
Launched in February, Microsoft Azure Machine Learning is a platform designed for processing massive amounts of data and building predictive applications. The Microsoft Azure ML platform provides capabilities such as natural language processing, recommendation engine, pattern recognition, computer vision, and predictive modeling.
The Microsoft Azure ML documentation contains a ton of information. However, information for many of the services is spread across different sections of the Azure website (and some information is on the Project Oxford website), making it somewhat hard to follow. There is an Azure Machine Learning Gallery where all of the machine learning APIs, experiments, and tutorials are listed in one place.
While the Microsoft Azure ML platform is rather new, the service has already gained significant popularity. It will be interesting to see how Microsoft's machine learning platform fares against Google, IBM, and Amazon in the coming months.
Amazon Machine Learning
Link: /api/amazon-machine-learningTrack this API
API Documentation URL: http://aws.amazon.com/documentation/machine-learning/
The Amazon Machine Learning platform has gained a lot of popularity in the short time since its launch in April. The service makes it possible to build intelligent applications that feature machine learning capabilities such as pattern recognition and prediction. Developers can use Amazon ML APIs to build applications that feature fraud detection, content personalization, document classification, customer churn prediction, and more.
Amazon provides comprehensive, detailed information about the Amazon ML platform and APIs. However, the documentation is somewhat hard to follow, and some of the information is provided in PDF format. The Amazon ML developer site features a large selection of SDKs and client libraries, a forum, an API reference section, machine learning concepts section, and more.
The Amazon ML service seems to be a bit more complicated than Google Prediction or Microsoft Azure ML. However, Amazon does provide visualization tools and wizards that help users with the process of creating machine learning models. Both Amazon ML and Microsoft Azure ML are new services that have become popular in a very short amount of time. It will be interesting to see which company, Microsoft or Amazon, will have the larger share of the machine learning cloud services market in the future.