Efficient deployment of predictive analytics through open standards and cloud computing

Abstract

Over the past decade, we have seen tremendous interest in the application of data mining and statistical algorithms, first in research and science and, more recently, across various industries. This has translated into the development of a myriad of solutions by the data mining community that today impact scientific and business applications alike. However, even in this scenario, interoperability and open standards still lack broader adoption among data miners and modelers. In this article we highlight the use of the Predictive Model Markup Language (PMML) standard, which allows for models to be easily exchanged between analytic applications. With a focus on interoperability and PMML, we also discuss here emerging trends in cloud computing and Software as a Service, which have already started to play a critical role in promoting a more effective implementation and widespread application of predictive models. As an illustration of how the benefits of open standards and cloud computing can be combined, we describe a predictive analytics scoring engine platform that leverages these elements to deliver an efficient deployment process for statistical models.