Python code for identification of localized energy and nonlinear waves in numerical simulations of one-dimensional crystal lattice models. Using the code and developed methodology, please cite Bajārs, J., Kozirevs, F.: Data-driven intrinsic localized mode detection and classification in one-dimensional crystal lattice model. Physics Letters A 436, 128071 (2022), DOI: 10.1016/j.physleta.2022.128071.
File LocalizedWaveIdentification_PureCode_WithoutData.zip contains pure code without precomputed data and images.
This research has been financially supported by the specific support objective activity 1.1.1.2. “Post-doctoral Research Aid” of the Republic of Latvia (Project No. 1.1.1.2/VIAA/4/20/617 “Data-Driven Nonlinear Wave Modelling”), funded by the European Regional Development Fund (project id. N. 1.1.1.2/16/I/001).
- To perform a numerical simulation of the lattice dynamics, run the file
main.py. - All the parameter values are saved in the dictionary param and set in the file
param_val.py. - All functions used during the numerical simulations, applications, and visualizations are defined in the folder
functions. - All images are saved in the folder
figures. - To collect different wave data from numerical simulations, run the file
collect_wave_data.py. - All collected wave data from numerical simulations is saved in the folder
saved_sim_data. - To obtain classification algorithms with dimensionality reduction algorithms PCA and LLE, run the files
classification_PCA.pyandclassification_LLE.py, respectively. - All trained classification algorithms are saved in the folder
saved_classifiers. - To test and obtain precision and recall scores, run the file
precision_recall_mean.py. - To identify nonlinear localized waves in numerical simulations using the built classification algorithms and sliding window approach, run the file
main_applications.py, where all the data is saved in the foldersaved_applications_data.