University of Houston

Houston and EMSL fold reaction chemistry into a neural network to forecast rock parameters

The framework is built so that a predicted mineral concentration cannot come out negative, and its real target is parameters no experiment supplies, which is also why the account leaves its accuracy unpriced.