GitHub - MathBioCU/WSINDy_PDE_OL: A weak-form algorithm for online discovery of PDEs from streaming data.

Branches Tags

Name		Name	Last commit message	Last commit date
Latest commit History 10 Commits
datasets		datasets
#get_proj_op.m#		#get_proj_op.m#
README.txt		README.txt
add_noise.m		add_noise.m
build_fcn_lib_tags.m		build_fcn_lib_tags.m
build_linear_system.m		build_linear_system.m
build_str_tags.m		build_str_tags.m
coarsen_data.m		coarsen_data.m
convNDfft.m		convNDfft.m
display_results.asv		display_results.asv
display_results.m		display_results.m
findcornerpts.asv		findcornerpts.asv
findcornerpts.m		findcornerpts.m
gen_noise.m		gen_noise.m
get_astep.m		get_astep.m
get_initial_weights.m		get_initial_weights.m
get_lambda.m		get_lambda.m
get_largest.m		get_largest.m
get_lib_OL.m		get_lib_OL.m
get_lib_tags.m		get_lib_tags.m
get_phi_handle.m		get_phi_handle.m
get_proj_op.m		get_proj_op.m
get_sparse_update.m		get_sparse_update.m
get_testfcn_weights.m		get_testfcn_weights.m
liblisttree.m		liblisttree.m
loaddata.m		loaddata.m
ode_data_gen.m		ode_data_gen.m
offline_computations.asv		offline_computations.asv
offline_computations.m		offline_computations.m
partitionNk.m		partitionNk.m
partitions.m		partitions.m
plot_learnedc.m		plot_learnedc.m
print_pde.m		print_pde.m
print_results_OL.m		print_results_OL.m
save_movie.m		save_movie.m
sweep.m		sweep.m
tags2axi.m		tags2axi.m
tpscore.m		tpscore.m
wnorm.m		wnorm.m
wsindy_OL_inputs.m		wsindy_OL_inputs.m
wsindy_OL_script.m		wsindy_OL_script.m

Repository files navigation

The code here can be used to reproduce the results in "Online Weak-Form Sparse Identification of Partial Differential Equations" - D. Messenger, E. Dall'Anese, D. Bortz. 

The file 'wsindy_OL_inputs.m' is used to set all hyperparameters and specify data to be loaded in, while the file 'wsindy_OL_script.m' runs the algorithm, outputting to the terminal specifications on the learned PDE as the online algorithm runs. 

Only the Kuramoto-Sivashinsky dataset is under the 100Mb github file size limit, but the other datasets (variable coefficient wave equation in 2D and linear wave equation in 3D) are available by request ([email protected]).