I am trying to achieve the best fit for a non linear model, yet I need to ensure that certain coefficients return as positive values. I have been looking for similar questions but I have found nothing at all.
Below is my code, and I need to ensure that "G, F and J" return as positive values, which doesn't happen for "G".
ClearAll[A, B, F, G, H, J, P, rb, nh, f, rp, mw, eff]
FIRL = Import[
"C:\\Users\\Utilizador\\Desktop\\TESE\\EXCELmathematica\\\
EffMathAdi - DEEPER - M2 - ATTEMPT BOTH MASSES SEMBP750W.xlsx"];
(*CASO - 2x --------------------------------------------------------------------------------------------------------------*)
nlm = NonlinearModelFit[
FIRL[[1]], {( (P)^A*(rb)^B*(nh)^F*(rp)^G*(1 + f)^H*(mw)^
J)*100}, {{A, 0.00000001}, {B, 0.00000001}, {F, 0.00000001}, {G,
0.00000001}, {H, 0.00000001}, {J, 0.00000001}}, {P, rb, nh, f,
rp, mw}, Method -> {"Automatic"}]
nlm["ParameterConfidenceIntervalTable"]
nlm[{"ParameterTable", "RSquared"}]
(*DRL2=DRL2/.\[VeryThinSpace]Extract [nlm["BestFitParameters"],{1}];*)
A = A /. Extract [nlm["BestFitParameters"], {1}];
B = B /. Extract [nlm["BestFitParameters"], {2}];
F = F /. Extract [nlm["BestFitParameters"], {3}];
G = G /. Extract [nlm["BestFitParameters"], {4}];
H = H /. Extract [nlm["BestFitParameters"], {5}];
J = J /. Extract [nlm["BestFitParameters"], {6}];
FI = Table[{FIRL[[1, i, 7]], (FIRL[[1, i, 1]])^A*(FIRL[[1, i, 2]])^
B*(FIRL[[1, i, 3]])^F*(FIRL[[1, i, 5]])^G*(1 + FIRL[[1, i, 4]])^
H*(FIRL[[1, i, 6]])^J*100}, {i, 1, Dimensions[FIRL][[2]]}];
FP = Table[{FIRL[[1, i, 7]], FIRL[[1, i, 7]]}, {i, 1,
Dimensions[FIRL][[2]]}];
Thanks for taking the time to consider my question, I am praying for a simple command that forces the constrain...
{( (P)^A*(rb)^B*(nh)^F*(rp)^G*(1 + f)^H*(mw)^ J)*100,G>0&&F>0&&J>0}as your model specification (or any other constraint you want to which is consistent with your starting values). (Also btw, /. can work on multiple variable at once, you dont have to Extract it for every single one). $\endgroup$