LocalizedWaveIdentification

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).

Instructions to run the code

• 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.py and classification_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 folder saved_applications_data.