Predicting model of I–V characteristics of quantum-confined GaAs nanotube: a machine learning and DFT-based combined framework

Continuous developments of machine learning algorithms have covered the various ways to analyze the atomistic structure and characteristics of quantum-confined nanostructures effectively. This work presents a machine learning model based on a regression fine tree algorithm to resolve the current–voltage characteristics model for GaAs nanotube during quantum confinement. The nanotube is 3.52 nm long and 3.61 nm wide. This paper presents predictive distributions of the current–voltage characteristic model with a sufficiently high level of confidence. This is a challenging task due to the backscattering effect of the quantum-confined nanostructures while the channel length is beyond the mean free path. Due to this quantum interference, it is difficult to predict the current–voltage characteristics correctly for quantum-confined nanostructures. Therefore, this machine learning approach helps to predict the model almost accurately with negligible erroneous values. This framework introduces a combined approach for both DFT and machine learning algorithms with lesser time cost and high predictivity response.