This project predicts flatbands from bandstructures in 2Dmatpedia database (http://www.2dmatpedia.org) using a supervised CNN and identifies isostructural groups among flatband compounds using a hybrid clustering algorithm.
The algorithms are divied into 6 modules (in 6 folders). They are numbered according to the sequence of application.
The modules are :
1.Trained_CNN_from_Materials_project
2.Processing_2Dmatepedia_downloaded_data_for_NN_model
3.Flatness_prediction_&_visualize_2Dmatpedia_BS_segments
4.Compound_flatness_calculation
5.Sublattice_extraction_and_vectorization
6.Clustering_module