Fractal Modeling of Biomedical Time Series in S-PLUS

Period of Performance: 04/25/2002 - 11/30/2002


Phase 1 SBIR

Recipient Firm

Insightful Corporation
Seattle, WA 98109
Principal Investigator


The goal of this project is to provide a comprehensive state-of-the-art software package, S+FRACTAL for modeling and data analysis of fractal processes. The overall implementation will be in the S-PLUS object oriented language. The specific aim of the software is to provide comprehensive and effective tools for estimating deterministic or stochastic fractal model parameters and to encourage research-oriented exploration of fractal structure found in biomedical time series. A diagnostic test suite will e implemented to infer the presence of deterministic or stochastic components in a time series and to aid the researcher in identifying an appropriate (fractal) model for arbitrary time series. Advanced linear and non-linear denoising techniques will be implemented to help separate the underlying deterministic dynamics from stochastic noise. Coupled with the existing statistical analysis capabilities of S-PLUS, the S+FRACTAL module will provide a commercially viable and comprehensive fractal analysis product for medical and health related fields as well as engineering disciplines. PROPOSED COMMERCIAL APPLICATIONS: Stochastic colored noise and non-linear deterministic chaotic fractal processes have emerged as important model for understanding complex physiological systems. There is currently no commercial software available for comprehensive analysis of fractal time series covering both non-linear deterministic chaotic and stochastic colored noise models. The proposed S+FRACTAL module will fill in this gap and allow MathSoft, Inc, to capture a significant edge in the market for biomedical time series analysis as well as extended markers in education, finance, and engineering.