#pmml

PMML -> predictive modeling markup language is the solution if you need to decouple modeling (producing/building) and prediction (consuming/scoring) parts. Imagine that you have worked in one environment let say used scikit learn for the model generation but would prefer to score your model in Spark Mlib or vice versa. For such cases, you need to define/export your model in a common format that can be used in prediction phase. Almost all of the machine learning libraries allow us to export our models in PMML (XML) and use this model in a different or same library. This is not only an interchange format, but it will also give us the flexibility of real-time model updates even if used in same environments.

Here is the pseudo like the simple line:

yourmodelObject.toPMML("/pathtofile")

Agile Deployment of Predictive Solutions using Zementis

Agile Deployment of Predictive Solutions using Zementis : Zementis, provides software solutions for predictive analytics. The core PMML based solutions are ADAPA, a decision engine for predictive

Agile Deployment of Predictive Solutions using Zementis

Agile Deployment of Predictive Solutions using Zementis : Zementis, provides software solutions for predictive analytics. The core PMML based solutions are ADAPA, a decision engine for predictive