Data Driven Design of Binary Metal Oxides Electrocatalyst for Oxidation / Evolution Reactionubmit This Sample Abstract Nirwan Syarif (a*), Herlina (b), Widia Purwaningrum (a)
a) Department of Chemistry, Universitas Sriwijaya
* nsyarif[at]unsri.ac.id
b) Department of Physics, Universitas Sriwijaya
Abstract
This paper reports on a predictive model we developed for catalyst design. This work began with collecting data on electrocatalysts prepared using transition metal oxides- this data set is then referred to as a dataset. The application of algorithms is crucial in learning to obtain high-performance classification and prediction models. The database machine learning used in this study is a support vector machine algorithm with parameters for the formation of the learning model being overpotential and current density- while the variable is the composition of the metal oxide. This work begins with collecting data on electrocatalysts prepared using transition metal oxides- this data set is then referred to as a dataset. The application of algorithms is crucial in learning to obtain high-performance classification and prediction models. To assess the model performance, a confusion matrix available in the Scikit-learn module is used.
Keywords: machine-learning- electrochemistry- chemi-informatics- material- energy
