Bios Data Analysis. Part 4. Flux and Action: Process Statistics

New methods for time series analysis describe patterns of variation and thereby identify creative phenomena. Regarding variation as meaningful change rather than meaningless variability, process statistics (1) analyzes time series; (2) measures changes in statistical parameters; (3) focuses on asymmetry and causality, the properties of directed action; and (4) rejects methodological assumptions and data transformations that prevent the identification of creative processes. Asymmetry and diversification (increase in variance with duration or embeddings) characterize many natural processes, mathematical bios and noise as creative, distinguishing them from random, periodic, and chaotic series. Partial autocorrelation characterizes empirical and mathematical biotic series as causal processes, differentiating them from stochastic series.