On the nature of mixed-type features in materials datasets

We provide four crystalline materials datasets that contain both numerical and categorical features of materials. The Lattice dataset [1,2] contains 1,439 binary AB body-centered cubic crystals, as described by 12 numerical and 9 categorical variables. The lattice constant is considered as the physical property of interest. The TC dataset [2,3] contains 101 binary alloys of transition and rare earth metals, as described by 15 numerical and 17 categorical features. The Curie temperature (TC) is considered as the property of interest. The Octet binary materials dataset [4] contains 82 materials, as described by 11 numerical and 2 categorical features. The difference in LDA energy between RS and ZB (∆E = E(RS) − E(ZB)) in eV is considered as the physical property of interest. The Fm3m dataset [5] contains 239 binary compounds collected from the Materials Project, as described by 12 numerical and 17 categorical variables. The formation of energy is considered as the physical property of interest. References: [1] K. Takahashi, L. Takahashi, J. D. Baran, and Y. Tanaka, "Descriptors for predicting the lattice constant of body centered cubic crystal", The Journal of chemical physics 146, 204104 (2017).<br> [2] D.-N. Nguyen, T.-L. Pham, V.-C. Nguyen, T.-D. Ho, T. Tran, K. Takahashi, and H.-C. Dam, "Committee machine that votes for similarity between materials", IUCrJ 5, 830-840 (2018).<br> [3] Y. Xu, M. Yamazaki, and P. Villars, "Inorganic materials database for exploring the nature of material", Japanese Journal of Applied Physics 50, 11RH02 (2011).<br> [4] L. M. Ghiringhelli, J. Vybiral, S. V. Levchenko, C. Draxl, and M. Scheer, "Big data of materials science: critical role of the descriptor", Physical review letters 114, 105503 (2015).<br> [5] A. Jain, S. P. Ong, G. Hautier, W. Chen, W. D. Richards, S. Dacek, S. Cholia, D. Gunter, D. Skinner, G. Ceder, et al., "Commentary: The materials project: A materials genome approach to accelerating materials innovation", Apl Materials 1, 011002 (2013).