Has anyone apply Kolmogorov Complexity to uncover hidden causal factors in price actions akin to what is described in the following paper?
Unsupervised network deconvolution is an emerging method for discovering underlying structures in graphs without prior knowledge of features or labels. The algorithm aims to discover a tree structure at some level of depth, with tree leaves closer to each other when such objects have a common or similar causal mechanism and for which no feature of interest has been selected. This is achieved by partitioning a graph into N components by removing edges that cause the least information loss.
https://www.nature.com/articles/s42256-018-0005-0
Thank you for any pointers.
Yeu