A NOVEL CRYSTAL STRUCTURE PREDICTION USING HYBRID METHOD

Crystal structure prediction is the most challenging problem for material science and the chemical field.Nowadays, there are many methods have been developed in machine learning and deep learning methods to predict the crystal structure with less computational time.In this paper, we combine the machine learning and deep learning methods to predict the crystal structure.We propose the Deep-Machine Learning methods to classify the crystal structure.In this work, to identify the four crystal structures.They are orthorhombic, cubic, tetragonal, and rhombohedral with the best accuracy.To harness the deep-machine learning methods to classify the crystal structure, it gives the best accuracy and less time complexity.Corr-ANB model gives the 87% accuracy of predict the crystal structure in 10-fold cross validation.Our method works very fast, and it is inexpensive.