michele2198

Thank you for answer and considering this work, I am in a project where we aim at comparing many combination of feature reduction + ML combinations so it would help...

Hi I managed to do this using an AMLPipeline: could you help me with the graph of feature importance? I am stuck with the Dataframes.by which should be now groupby!...

No worry, I made it work with classification, see above example, and also made the graph (I will post it in a few hours). Thanks best regards

The graph: ``` data_plot = DataFrames.groupby(data_shap, [:feature_name]); baseline = round(data_shap.intercept[1], digits = 1) using DataFramesMeta regr=sort(@combine(data_plot,:x2 = sum(:shap_effect)),order(:2,rev=true)); using Gadfly # Plotting p = Gadfly.plot(regr[1:10,:], y = :feature_name, x =...

Sure, here is all the code: ``` `using Pkg; Pkg.add("XLSX") using XLSX import Base.Threads.@threads #Pkg.add("MLDataPattern") #Pkg.add("OutlierDetection") #Pkg.add("OutlierDetectionData") using MLDataPattern #X_bal, y_bal = oversample((X, y)); #using OutlierDetection #using OutlierDetectionData: ODDS using...

no, consider that shapley is not designed for binary class, but for regression. So I don't see reason to not work with multiclass my code is designed to predict the...