YPS: Yottamine Predictive ServicesSVM, Machine Learning in the Amazon Cloud

YPS is a family of high performance, cloud-based, low-cost software services for data scientists and predictive modelers. YPS provides an R integration package, a Java API, and a web based interface.

SAN DIEGO, CA, 02/20/13 - Seattle-based
YottamineYottamine Analytics, a pioneer in high performance prediction, announced the availability of Yottamine Predictive Services (YPS), a family of cloud-based software services that enable data scientists to rapidly build highly accurate predictive models on-demand, with less trial and error, and at the lowest possible cost.

The company also announced the availability of YottamineR®, a free R language programming interface for Yottamine's growing family of predictive services.

YPS is available exclusively to users of Amazon's AWS Elastic Compute Cloud (EC2) and Simple Storage Service (S3). These Amazon services allow YPS users to "rent" just the amount of computing power and data storage they need to build new predictive models using input data of any size, from business and scientific information to Big Data from the Web, machines, and sensor networks. YPS produces robust, efficient models that can be used to score very large data sets on AWS or on other systems.

In addition to being fast, accurate and scalable, Yottamine's predictive modeling services are also highly automated. YPS suggests optimal parameter values, saving hours of tedious tuning, and the software automatically calculates and provisions AWS resources, eliminating system administration chores.

YPS uses highly parallel support vector machine (SVM) algorithms that are optimized for running on scalable cloud computing clusters and for working with Big Data. YPS currently includes four different algorithms for solving different kinds of problems - Linear, Gaussian, Polynomial, and Nearest Neighbor.

The decision boundary of a Nearest Neighbor SVM for the Two Moon problem

YPS services are priced from $10 to $50 per hour, per compute node, depending on algorithm used. For a limited time, new users will receive their first month free and second month for half off.

Data scientists can access all of Yottamine's services using the R statistical programming language with YottamineR®, a free downloadable R programming package. It enables R programmers to easily build, score and visualize new predictive models using familiar tools, and to incorporate Yottamine's predictive services into existing workflows. Please visit www.yottamine.com/YottamineR.html for a demo.

YPS also provides a Java API, and a web based interface.

For more information, please visit yottamine.com.