Using Imprecise Complexes To Computationally Recognize Causal Relations In Very Large Data Sets

Lawrence J. Mazlack

Computationally recognizing causal relationships in data is fundamentally important to good decision-making. There are vast amounts of computer stored, multi-faceted data. Understanding how stored data items affect each other is crucial in making good decisions. The most important decisional information is an understanding of causal relationships. A method to discover causality from large, observational data sets would be transformative. Broadly effective computational methods are computationally untested. An abundance of digital data riches promise a profound impact in both the quality and rate of discovery and innovation in science and engineering, as well as in other societal contexts. Worldwide, researchers are producing, accessing, analyzing, integrating and storing massive amounts of digital data daily, through observation, experimentation and simulation, as well as through the creation of collections of digital representations of tangible artifacts and specimens. Modern experimental and observational instruments generate and collect large sets of data of varying types (numerical, video, audio, textual, multi-modal, multi-level, multi-resolution) at increasing speeds. Often, the data users are not the data producers, and they thus face challenges in harnessing data in unforeseen and unplanned ways. In many science or engineering applications, for example, in mesoscale weather prediction or critical infrastructure protection applications, the ability to gather, organize, analyze, model, and visualize large, multi-scale, heterogeneous data sets in rapid fashion is often crucial.

PDF full paper