Level: Intermediate

Author(s): James D. Whiteside, II PE FAACE; Gregory J. Whiteside, PE CCP

Venue: 2017 AACE International Annual Meeting, Orlando, FL

Abstract: Standard practices for statistical and probability analysis are not good applications for small data populations. Two better tools are introduced for small data population analysis: the discrete Fourier transform and phase space to perform statistical and probability analysis. Rather than rely on estimating practices (statistical and probability analysis), phase space and Fourier transform are objective data-driven empirical analyses to directly compute probability distributions, model data, forecast and evaluate for cyclical responses.

Phase space analysis is an empirical solution for data modeling, risk assessment, statistical analysis and an alternative tool to multivariate regression. People are pre-disposed (biased) to look for coincidences. The basis of estimating practices (statistics and probability) assumes data is stochastic (random) when, in fact, data is engineered, collected and analyzed by people who are behavioral. If a cyclic condition is determined to exist, then the root cause or trend may be behavioral.

Author(s): James D. Whiteside, II PE FAACE; Gregory J. Whiteside, PE CCP

Venue: 2017 AACE International Annual Meeting, Orlando, FL

Abstract: Standard practices for statistical and probability analysis are not good applications for small data populations. Two better tools are introduced for small data population analysis: the discrete Fourier transform and phase space to perform statistical and probability analysis. Rather than rely on estimating practices (statistical and probability analysis), phase space and Fourier transform are objective data-driven empirical analyses to directly compute probability distributions, model data, forecast and evaluate for cyclical responses.

Phase space analysis is an empirical solution for data modeling, risk assessment, statistical analysis and an alternative tool to multivariate regression. People are pre-disposed (biased) to look for coincidences. The basis of estimating practices (statistics and probability) assumes data is stochastic (random) when, in fact, data is engineered, collected and analyzed by people who are behavioral. If a cyclic condition is determined to exist, then the root cause or trend may be behavioral.