English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

Progress in the application of machine learning (ML) methods to materialsdesign is hindered by the lack of understanding of the reliability of MLpredictions, in particular for the application of ML to small data sets oftenfound in materials science. Using ML prediction for transparent conductor oxideformation energy and band gap, dilute solute diffusion, and perovskiteformation energy, band gap and lattice parameter as examples, we demonstratethat 1) analysis of ML results by construction of a convex hull in featurespace that encloses accurately predicted systems can be used to identifyregions in feature space for which ML predictions are highly reliable 2)analysis of the systems enclosed by the convex hull can be used to extractphysical understanding and 3) materials that satisfy all well-known chemicaland physical principles that make a material physically reasonable are likelyto be similar and show strong relationships between the properties of interestand the standard features used in ML. We also show that similar to thecomposition-structure-property relationships, inclusion in the ML training dataset of materials from classes with different chemical properties will not bebeneficial and will slightly decrease the accuracy of ML prediction and thatreliable results likely will be obtained by ML model for narrow classes ofsimilar materials even in the case where the ML model will show large errors onthe dataset consisting of several classes of materials. Our work suggests thatanalysis of the error distributions of ML predictions will be beneficial forthe further development of the application of ML methods in material science.

Identification of high-reliability regions of machine learning  predictions in materials science using transparent conducting oxides and  perovskites as examples