ニューラルネットワーク計算による2元系固体化合物の標準エントロピーの推算

抄録

In the present study, a layer type neural network computation was applied to estimate the standard entropy of binary solid oxides, sulfides and halides. Independent variables to influence the thermodynamics property associated with dispersion or randomness in the crystals were used as input parameters for the calculation. 325 substances involving 12 input parameters were applied to the calculation. The regression computation enabled reproduction of training data cited in learning process and prediction of test data not used in the learning process with high accuracy.

In addition, the contribution of each input property to the estimation of the standard entropy was also evaluated. It was found that the volume and the weight per a composition had positive impacts, and the atomic weight and the orbital radius of an anion had negative impacts. Furthermore, it was suggested that coordination numbers of composition elements have little effect on the precision of reproduction and prediction of the standard entropy.

Fig. 5　Comparison between estimated value by neural network computation with assessed input parameters and recommended value in literature in each regression.