Machine Learning modelling the IAEA stopping power database

D.M. MITNIK; F. BIVORT HAIEK; A. M. P. MENDEZ; C. C. MONTANARI

Helsinki

Conferencia; 29th international conference on atomic collisions in solids & 11th international symposium on swift heavy ions in matter; 2022

Helsinki University

The International Atomic Energy Agency (IAEA) stopping power database [1] is a highly valuable public resource, making available compilations of the majority of the experimental measurements published over the last nine decades. The database is continuously updated, and being accessible to the global scientific community. Our purpose is to extend the database, allowing to apply machine learning methods to predict the electronic stopping power cross section for any ion and target combination for a wide range of incident energies.The IAEA database was build gathering published articles by diverse authors. Hence, it is not standarised in its original form and in many cases different experiments are not consistent to one another. The data may have been presented in distinct units and formats. We devoted a significant effort in the reorganization of the database, unifying units, and arranging the data in a standard (csv) format, enabling the obtention of the information to be easily and quicklyaccesible. Considering that the database contains several dozen of thousands input values, purging manually this data requires also a considerable amount of work. To this purpose, we developed a machine learning method to clean up the database, based on an unsupervised clustering technique (DBSCAN) which identifies outlier values, and determines which data to keep in cases of overlapping, taking into account the year of production. The refined data is used to train a deep neural network, able to accurately reproduce the input experimental results, and to predict new results in cases which have never been seen in the training procedure (the test set). The code ESPNN (Electronic Stopping Power with Neural Network), performing all these mentioned tasks, is presented in this work.