Automatic model generation
For chemists, this automates the tedious process of mechanism selection, but the method is demonstrated only on small-scale examples.
The paper proposes a method to automatically generate and select chemical reaction mechanisms that best fit experimental data, demonstrated on an artificial example and a small real dataset.
The goal of the paper is to automatize the selection of mechanisms which are able to describe a set of measurements. In order to do so first we construct a set of possible mechanism fulfilling chemically reasonable requirements with a given number of species and reaction steps. Then we try to fit all the mechanisms, and offer the best fitting one to the chemist for further analysis. The method can also be used to a kind of lumping: to reproduce the results of a big mechanism using a smaller one, with less number of species. We show two applications: one on an artificial example and another one on a small real life data.