The High Dimensional Model Representation (HDMR) provides a physically intuitive method for exploring the relations among input variables and between inputs and outputs for complex systems. ExploreHD^TM provides a simple and intuitive interface for studying these relations. HDMR expands the response function of a system into a heirarchical representation, consisting of one, two and three parameter functions of the inputs. The method produces a Fully Equivalent Operational Model (FEOM) which captures the significant cooperative relations between the input variables.

ExploreHD^TM provides a window into your data allowing you to quantify the global sensitivity of your system to each variable and allowing you to determine which are the most significant.

Available for Windows, Linux and MacOS X Generate HDMR models to 1st, 2nd or 3rd Order Perform global sensitivity analysis Produce Fully Equivalent Operational Model Generate optimal model reductions Expansion in optimal orthogonal polynomials Ability to randomly divide data into training and test sets User selectable confidence level Generate random distributions of inputs for models Visualize 1st and 2nd order function surfaces Visualize model truth plot Import Text Filter Export data to CSV files

Some Typical Uses

HDMR has been used to

• Reduce complex chemical reaction networks by determining the significant species contributors
• Simulate bioremediation of uranium contaminanted aquifiers
• Discover new materials in multicomponent semiconductor physics

• For more information contact: Dr. Luwi Oluwole