What happened?
guesses gets confused with "im" if dealing with complex numbers. Example input:
Complexity Loss Equation
1 0.060197003 RC_vector
3 0.017789923 (-0.2052275f0 - 0.016992446f0im) + RC_vector
5 0.017055396 (-0.21000202f0 - 0.016444953f0im) + (RC_vector / (0.97645104f0 + 0.00017897492f0im))
guesses = CSV.read("...\\hall_of_fame.csv", DataFrame).Equation
model = SRRegressor(
niterations = 1_000_000,
binary_operators = [+, -, *, /],
maxsize = 35,
guesses = guesses,
parallelism = :multiprocessing,
procs = procs,
populations = 32,
population_size = 200,
batch_size = 5000,
batching = true,
parsimony = 0.0005,
)
mach = machine(model, X, y)
fit!(mach)
Code will run fine if I comment out guesses.
Thanks again for your support of this amazing package!
Version
SymbolicRegression v2.0.0-alpha.9
Operating System
Windows
Interface
Julia REPL
Relevant log output
[ Info: It seems an upstream node in a learning network is providing data of incompatible scitype. See above.
ERROR: ArgumentError: Variable `im` not found in `variable_names`. Consider interpolating with $ if passing a value.
Stacktrace:
[1] parse_leaf(ex::Symbol, variable_names::Vector{String}, node_type::Type{DynamicExpressions.NodeModule.Node{ComplexF32, 2}}, expression_type::Type{Expression}; kws::@Kwargs{})
@ DynamicExpressions.ParseModule C:\Users\George\.julia\packages\DynamicExpressions\x2aIa\src\Parse.jl:453
[2] parse_leaf
@ C:\Users\George\.julia\packages\DynamicExpressions\x2aIa\src\Parse.jl:433 [inlined]
[3] #_parse_expression#19
@ C:\Users\George\.julia\packages\DynamicExpressions\x2aIa\src\Parse.jl:430 [inlined]
[4] _parse_expression
@ C:\Users\George\.julia\packages\DynamicExpressions\x2aIa\src\Parse.jl:421 [inlined]
[5] #12
@ C:\Users\George\.julia\packages\DynamicExpressions\x2aIa\src\Parse.jl:345 [inlined]
[6] map(f::DynamicExpressions.ParseModule.var"#12#16"{Expression, DynamicExpressions.NodeModule.Node{ComplexF32, 2}, @Kwargs{}, OperatorEnum{Tuple{Tuple{}, Tuple{typeof(+), typeof(-), typeof(*), typeof(/)}}}, Vector{String}, Nothing}, t::Tuple{Float32, Symbol})
@ Base .\tuple.jl:356
Extra Info
No response
What happened?
guesses gets confused with "im" if dealing with complex numbers. Example input:
Complexity Loss Equation
1 0.060197003 RC_vector
3 0.017789923 (-0.2052275f0 - 0.016992446f0im) + RC_vector
5 0.017055396 (-0.21000202f0 - 0.016444953f0im) + (RC_vector / (0.97645104f0 + 0.00017897492f0im))
Code will run fine if I comment out guesses.
Thanks again for your support of this amazing package!
Version
SymbolicRegression v2.0.0-alpha.9
Operating System
Windows
Interface
Julia REPL
Relevant log output
Extra Info
No response