m3g / easyfit.jl Goto Github PK

Hi,
Thank you for this very nice package!
I noticed that the exponential fit is sometimes unstable and converges to nonsense values (screenshots below).

I did not investigate further, but if needed I'd be happy to.
Best,
Pierre

@pat-alt

I merged the PR and did some code refactoring. But the polynomial fits do not seem to be working. I currently have no time to debug it, but it seems to be something related to how the model is being built. I would suggests following the same structure as the multi-exponential functions.

The tests are now in the end of the corresponding file of each type of fit. You can run them as usual, but also individually if using VScode.

Current Julia versions (1.8) often build dependency trees requiring Interpolations.jl v0.14 and newer.
Would be nice to add this and trigger a release. The commands that I use (like movavg) seem wo work also with the newer Interpolations.jl version, however I did not check every function.

PS: If you're lazy I opened a PR :-)

Neat package! Not sure if you have plans for future developments, but one thing that would be nice to have is the ability to predict from the fitted object, e.g.:

function (fit::EasyFit.Cubic)(x::Real)
    a = fit.a
    b = fit.b
    c = fit.c
    d = fit.d
    return a * x^3 + b * x^2 + c * x + d
end

Then you can just call it on new data:

julia> x = sort(rand(10)); y = x.^3 .+ rand(10);

julia> f = fitcubic(x,y)

 ------------------- Cubic Fit ----------------- 

 Equation: y = ax^3 + bx^2 + cx + d 

 With: a = 3.498571133673037
       b = -5.75292789995513
       c = 2.626129810011887
       d = 0.6361773562878126

 Pearson correlation coefficient, R = 0.7405690253097572
 Average square residue = 0.01215483592609077

 Predicted Y: ypred = [0.6416314330095221, 0.6417874373639705...
 residues = [-0.13182717628179608, -0.01592993507117535...

 ----------------------------------------------- 


julia> f.(rand(10))
10-element Vector{Float64}:
 0.8761239348448231
 0.9115358893542463
 0.9121562305431836
 0.8919530945018805
 ⋮
 0.81693749334824
 0.9622975666245418
 0.9753695182250022

This issue is used to trigger TagBot; feel free to unsubscribe.

If you haven't already, you should update your TagBot.yml to include issue comment triggers.
Please see this post on Discourse for instructions and more details.

If you'd like for me to do this for you, comment TagBot fix on this issue.
I'll open a PR within a few hours, please be patient!