Reduce data sets to a more manageable size, without compromising interpretability, for quicker machine learning model training and deployment
CHICAGO – InRule Technology®, an intelligence automation company providing integrated decisioning, machine learning and process automation software to the enterprise, today announced that its suite of machine learning modeling engines now includes integrated downsampling and model calibration, allowing for faster model creation without sacrificing the interpretability of model outputs.
InRule’s downsampling capability reduces the overall size of a data set, allowing teams to train machine learning (ML) models more quickly when working with data sets where the prediction target rarely occurs, while reducing cloud costs for Model Ops infrastructure.
Complementary to integrated downsampling, InRule Machine Learning users can now leverage new model calibration functionality prior to model deployment. Incorporating model calibration allows users to re-align model confidence outputs to the rate of real-world occurrence, improving actionability and interpretability by adjusting the distortion from downsamping data sets.
According to Danny Shayman, InRule Machine Learning product manager, “Many of our customers’ most common use cases involve highly class-imbalanced data sets and this pair of capabilities will significantly reduce their model training times while retaining full interpretability.”
# # #
About InRule Technology
InRule Technology® is an intelligence automation company providing integrated decisioning, machine learning and process automation software to the enterprise. By enabling IT and business leaders to make better decisions faster, operationalize machine learning and improve complex processes, the InRule® Intelligence Automation Platform increases productivity, drives revenue, and provides exceptional business outcomes. More than 500 organizations worldwide rely on InRule for mission critical applications. InRule Technology has been delivering measurable business and IT results since 2002. Learn how to make automation accessible at www.inrule.com.
InRule and InRule Technology are registered trademarks of InRule Technology, Inc.