An efficient method of searching for correlations between unlimited datasets to provide forecasting models

This work presents a highly powerful and very simple statistical method for creating predictive models based on structure–property relationships. The algorithm was implemented in the software CORRELATO, which demo version is presented in Online Supplementary Materials. This algorithm was tested on a small series of phthalocyanine dyes to search for the relationship between optical limiting effect and quantum chemical descriptors responsible for the nonlinear optical properties of these absorbers