"Dataset.mat" provides the dataset for training and test.
"modelGen.m" is used for training and evaluating the NN model. It is the code for Fig. 4.
"Multi_BMG.xlsx" includes the data of multicomponent MGs and crystals, and BMG data with maximum formation sizes.
"calSpecComp.m" gives the predictions for multicomponent alloys, and generates Fig. 9 and 10.
"findSpec.m" gives the predictions for a ternary alloy system, such as Fig. 5.
"EleIndex.max" index of elements
"prec_rec.m" code for calculating precision-recall curve. It is downloaded from the link "https://www.mathworks.com/matlabcentral/fileexchange/21528- % precision-recall-and-roc-curves"
"NN93.m" the NN model
"NN93_new.mat" the parameters and performance of the NN model, which is shown in Fig. 4.
"fullFeatures.mat" database of ternary alloys, which includes the data of 3,745,665 compositions in 16,215 systems. It is generated by "createFullFeature.m"
"properties20190429.xlsx" is the properties of elements.
"mixingenthalpy20190328.xlsx" is the form of mixing enthalpy.
"normx.m" nomalization function.