Arthur A Jamshidi

Title: Arthur A Jamshidi: Innovator in Nonlinear Function Approximation

Introduction

Arthur A Jamshidi is a notable inventor based in Plainsboro, NJ (US). He has made significant contributions to the field of nonlinear function approximation, holding 2 patents that showcase his innovative approach to constructing models from high-dimensional data.

Latest Patents

One of his latest patents involves an algorithm for nonlinear function approximation over high-dimensional domains. This algorithm is designed to construct nonlinear models from scattered data, progressing iteratively by adding new basis functions to refine the model. The placement of these basis functions is determined by a statistical hypothesis test that reveals geometric structures when it fails. Each added function is fitted to data within a defined local region, allowing for the determination of parameters, particularly the scale of the local model. Notably, this method does not require ad hoc parameters, as the algorithm automatically determines the number of basis functions needed for an accurate fit. This approach can be applied to various problems, including modeling data on manifolds and predicting financial time-series. The algorithm is primarily presented in the context of radial basis functions but can also be utilized with other function approximation methods, such as multi-layer perceptrons.

Career Highlights

Throughout his career, Arthur has worked with esteemed organizations, including the National Science Foundation and the Colorado State University Research Foundation. His work has significantly impacted the field of data modeling and function approximation.

Collaborations

Arthur has collaborated with notable individuals in his field, including Michael Joseph Kirby, enhancing the depth and breadth of his research and innovations.

Conclusion

Arthur A Jamshidi's contributions to nonlinear function approximation demonstrate his innovative spirit and dedication to advancing technology. His work continues to influence various applications in data modeling and prediction.