Material structure-property prediction using orthogonal functional basis neural network

An important trend in materials research is to predict properties for a new material. Often the prediction is motivated by the search for a material with several important materials property features. The selection of the property features is crucial to the plausibility of the prediction. This paper proposes a neural-network computing approach to evaluate this issue. With the proposed approach, we are able to predict property features for an unknown compound. In this paper we summarize the prediction attained with the proposed neural network structure-orthogonal functional basis neural network (OFBNN). The network, which combines a new basis selection process and a regularization technique, not only gives us a more computationally tractable method, but better generalization performance. Simulation studies presented demonstrate the performance, behavior, and advantages of the proposed network.