Edit on GitHub

Solve

All algorithms have been combined under a single API to match the interface of other SciML packages. Thus, you can simply define a Problem, and then seamlessly switch between solvers.

The table below provides an overview, which class of algorithms support which class of problems.

DirectDiscreteContinuousBasisRequires
Koopman-++Optional
Sparse Regression+++Necessary
EQSearch+++NoSymbolicRegression.jl
OccamNet+++NoFlux.jl

All of the above methods return a DataDrivenSolution if not enforced otherwise.

Common Options

Many of the algorithms implemented directly in DataDrivenDiffEq share common options. These can be passed into the solve call via keyworded arguments and get collected into the CommonOptions struct, which is given below.

DataDrivenDiffEq.DataDrivenCommonOptions — Type
mutable struct DataDrivenCommonOptions{T, K}

Common options for all methods provided via DataDrivenDiffEq.

Fields

  • maxiter

    Maximum iterations

  • abstol

    Absolute tolerance

  • reltol

    Relative tolerance

  • progress

    Show a progress

  • verbose

    Display log - Not implemented right now

  • denoise

    Denoise the data using singular value decomposition

  • normalize

    Normalize the data

  • sampler

    Sample options, see DataSampler

  • f

    Mapping from the candidate solution of a problem to features used for pareto analysis

  • g

    Scalarization of the features for a candidate solution

  • digits

    Significant digits for the parameters - used for rounding. Default = 10

  • kwargs

    Additional kwargs

source
Info

The keyword argument eval_expression controls the function creation behavior. eval_expression=true means that eval is used, so normal world-age behavior applies (i.e. the functions cannot be called from the function that generates them). If eval_expression=false, then construction via GeneralizedGenerated.jl is utilized to allow for same world-age evaluation. However, this can cause Julia to segfault on sufficiently large basis functions. By default eval_expression=false.

Solving the Problem

After defining a problem, we choose a method to solve it. Depending on the input arguments and the type of problem, the function will return a result derived the algorithm of choice. Different options can be provided, depending on the inference method, for options like rounding, normalization, or the progress bar. A Basis can be used for lifting the measurements.

# Use a Koopman based inference
res = solve(problem, DMDSVD(), kwargs...)
# Use a sparse identification
res = solve(problem, basis, STLQS(), kwargs...)

The DataDrivenSolution res contains a result which is the inferred system and a Basis, metrics which is a NamedTuple containing different metrics of the inferred system. These can be accessed via:

# The inferred system
system = result(res)
# The metrics
m = metrics(res)

Since the inferred system is a parametrized equation, the corresponding parameters can be accessed and returned via

# Vector
ps = parameters(res)
# Parameter map
ps = parameter_map(res)