Roy Debarati Dey, Roy Pradipta, De Debashis
Department of Electronics & Communication Engineering, B. P. Poddar Institute of Management & Technology, 137, V. I. P. Road, Kolkata, West Bengal, 700052, India.
Department of Computer Science & Engg., Maulana Abul Kalam Azad University of Technology, NH-12(Old NH-34), Haringhata, Post Office-Simhat, P.S.-Haringhata, Nadia, West Bengal, 741249, India.
J Mol Model. 2023 Oct 13;29(11):338. doi: 10.1007/s00894-023-05741-y.
The Adenine-based nanotube is theoretically designed, and its transmission spectra are investigated. The quantum-confined Adenine nanotube shows electronic transmission of the carrier at minimum stress. In this paper, the prediction of transmission spectra of the quantum-confined bio-molecular nanotube is investigated and deeply studied. Molecular level structure prediction and their electronic characterization can be possible with ab initio accuracy using a machine learning algorithmic approach. At the molecular level, it is difficult to predict quantum transmission spectra as these results are always hampered by the carrier backscattering effect. However, mostly these predictive models are available for intrinsic semi-conducting materials and other inorganic structures.
Machine learning algorithms are designed to predict the electronic properties of the nano-scale structure. This task is even more difficult when quantum-confined molecular arrangements are considered, whose transmission spectra are sensitive to the confinements applied. This paper presents an effective machine learning algorithms framework for predicting transmission spectra of quantum-confined nanotubes from their geometries. In this paper, we consider regression machine learning algorithms to find maximum accuracy with varying configurations and geometries to excerpt their atoms' local environment information. The Hamiltonian components are then used to enable the utilization of the information to predict the electronic structure at any arbitrary sampling point or k-point. The theoretical basics introduced in this process help to capture and incorporate minor changes in quantum confinements into transmission spectra and provide the framework algorithm with more accuracy. This paper shows the ability to predict the accurate algorithmic models of the Adenine nanotube. In this framework, we have considered a tiny data set to achieve a rapid and reliable method for electronic structure determination and also propose the best algorithm for predictive model analysis.
基于腺嘌呤的纳米管是通过理论设计的，并对其透射光谱进行了研究。量子限制的腺嘌呤纳米管在最小应力下显示出载流子的电子传输。本文对量子限制生物分子纳米管的透射光谱预测进行了研究和深入探讨。利用机器学习算法方法可以从头算精度实现分子水平结构预测及其电子表征。在分子水平上，很难预测量子透射光谱，因为这些结果总是受到载流子背散射效应的阻碍。然而，大多数这些预测模型适用于本征半导体材料和其他无机结构。
设计机器学习算法来预测纳米尺度结构的电子性质。当考虑量子限制的分子排列时，这项任务更加困难，因为其透射光谱对所施加的限制很敏感。本文提出了一种有效的机器学习算法框架，用于从量子限制纳米管的几何结构预测其透射光谱。在本文中，我们考虑回归机器学习算法，以通过不同的构型和几何结构找到最大精度，以提取其原子的局部环境信息。然后使用哈密顿量分量来利用这些信息预测任意采样点或k点处的电子结构。在此过程中引入的理论基础有助于捕捉量子限制中的微小变化并将其纳入透射光谱，并为框架算法提供更高的精度。本文展示了预测腺嘌呤纳米管精确算法模型的能力。在这个框架中，我们考虑了一个小数据集，以实现一种快速可靠的电子结构确定方法，并提出了用于预测模型分析的最佳算法。
