Abstract
The modern machine learning methods allow one to obtain the data-drivenmodels in various ways. However, the more complex the model is, the harder itis to interpret. In the paper, we describe the algorithm for the mathematicalequations discovery from the given observations data. The algorithm combinesgenetic programming with the sparse regression. This algorithm allows obtaining different forms of the resulting models. Asan example, it could be used for governing analytical equation discovery aswell as for partial differential equations (PDE) discovery. The main idea is to collect a bag of the building blocks (it may be simplefunctions or their derivatives of arbitrary order) and consequently take themfrom the bag to create combinations, which will represent terms of the finalequation. The selected terms pass to the evolutionary algorithm, which is usedto evolve the selection. The evolutionary steps are combined with the sparseregression to pick only the significant terms. As a result, we obtain a shortand interpretable expression that describes the physical process that liesbeyond the data. In the paper, two examples of the algorithm application are described: thePDE discovery for the metocean processes and the function discovery for theacoustics.