Empirical Dynamic Modeling

Empirical Dynamic Modeling

Time series are serial projectoins of a variable from motion on an attractor
Time series are serial projections of a variable from motion on an attractor

To understand complex ecological systems, the Sugihara lab has come up with an equation-free approach. Equations used to make predictions in ecology have repeatedly been unreliable mainly because species have nonlinear interactions.

Instead, the Sugihara lab introduced the empirical dynamic modeling (EDM) based on the chaos theory. EDM makes it possible to predict behavior with limited behavior by abandoning the use of equations and instead using a data-driven model. The EDM uses only raw data as input & reveals hidden causal relationships.

EDM is a method that can be used beyond ecology; it is already being used in areas such as finances and neuroscience. EDM is a reliable and effective method, so Sugihara and his team envision this to become more broadly used, breaking the convention of relying on equations.

Read more about this on EDM Wiki and PNAS.