A Marriage of Data Models: Atigeo Enhances Its Hybrid Analytics

Insights, Jul 27, 2016

A Marriage of Data Models: Atigeo Enhances Its Hybrid Analytics

Imagine data science could help an enterprise business predict the demand for products internally and at various suppliers, and then use this information to determine the right times to purchase specific products from specific suppliers. You can stop fantasizing because this technology exists; for five years, a company named ClearSight Systems—now retired—used science like this to help businesses minimize costs, better allocate supplies, and avoid penalties for unsatisfied demands.

In the second quarter of this year, Atigeo acquired the intellectual property (IP) for this type of dynamic forecasting and other ClearSight technology, including a general purpose optimizer that makes our optimization algorithm customizable for different problems across multiple industries. One of those industries is healthcare, where the incorporation of the ClearSight IP will transform the way xPatterns and its healthcare solution evolve. With this acquisition, xPatterns will have its own built-in repair engine, one that assesses historical data about inferior decisions (e.g., initial miscounts, theorem failures, etc.), zeroes in on faults within the system, and automatically tunes and corrects either by replacing or modifying terms.

Atigeo Chief Scientist Wolf Kohn, is the original architect of the technology, whose application in the software, financial, and federal government sectors has delivered dramatic results. At a major enterprise business that specializes in supply chain enterprise resource planning (ERP) for products with fast tracks to obsolescence, the technology’s dynamic feedback approach delivered a cost ratio of 96.6 compared to costs incurred through the static allocation of resources. In short, the technology brought a 97 percent reduction in costs.

The underlying approach Kohn spearheaded is called Multiple Agent Hybrid Control Architecture (MA-HCA, formerly Multiple Agent Declarative Control Architecture). It earned its “hybrid” title by embedding high-level rule-based models (e.g., diagnosis and patient management in the area of healthcare) in low-level, continuous evolution models (e.g., pharmacology, quantitative pathophysiology, immunology) to build automatons, or control mechanisms, that comply with rule-based and continuum restraints. Historically, rule-based and evolution models have been incompatible, even dichotomous. This is a problem in healthcare, where, for example, many diagnostic images such as X-rays (rule-based models) are not in computer-readable form (evolution models).

MA-HCA makes the marriage of these models possible. It also makes it faster. The acquired IP includes adaptation, a component of MA-HCA that improves performance through the observation and response to a control system’s behavior in real time—all without halting or otherwise compromising the system’s function as a whole.

Enhanced data repair and forecasting, each integral components of MA-HCA, will be built into the xPatterns 7.0 platform in 2017. Together, these capabilities help deliver efficiency without compromise, optimizing the fused intelligence, rapid and scalable growth, and precise forecasting that comprise the foundation of xPatterns.

Put xPatterns to work for your organization. Contact our sales team to learn more about the insights xPatterns can deliver you: sales@atigeo.com.

Written by Sean Marshall