You are here

Deploying Advanced Analytics Using R & PMML


The amount of data being processed by organizations worldwide is only increasing. The New York Stock Exchange, for example, generates a terabyte of new trading data daily. Global Internet traffic will reportedly more than quadruple to 767 exabytes in three years.

Allowing this data to remain unused and merely stored is wasteful. Data is increasingly being treated as a perishable asset, which is spurring business interest in predictive analytics.

The challenge for organizations has been a lack of technologies that can help programmers quickly and easily develop projects. A key in changing the status quo has been the creation of Predictive Modeling Markup Language (PMML), a standard language for creating statistical and data-mining models. It allows a model to be developed using one system, and deployed on another system using a completely different application. Prior to the development of PMML, model conversion was a time-consuming, expensive process.