论文部分内容阅读
The machine-learning prediction of perovskite ABX3 and A2BBX6 based on the geometric predictor(tolerance factor t)can distinguish an average of 74%of the classification accuracy(even worse for halide perovskite),which can be improved over 80%if considering the octahedral factor(μ = rB/rX)and atomic packing fraction(η = Vatoms/Vcell).