shapley | Compute Shapley-Shorrocks value decompositions | Machine Learning library

I'm using R package GameTheory to calculate Shapley-Shubik power indices. The command itself is very simple, ShapleyShubik(quota, y, Names = NULL)

where quota is the minimum amount of votes to pass a vote, y seats of each party (number of) and Names are labels for the parties. It is simple to use 'manually', but I would like extend my usage to automate it to iterate through a vast amount of data that is compiled in an dataframe DF.

My dataframe DF includes four columns: AREA, PARTY LABEL, PARTYSEATS and MAJORITY:

I have a Geopandas df 'districts' of all districts in Paris and a shapley point object 'eiffel_tower' of the Eiffel Tower. When I execute

I have tuned a model using GridSearchCV. Now I would like to calculate the Shapley values and visualize them. The difficulty is that the shap package excepts a model, not the GridSearch Results. Likewise it does not like when I pass it the best_estimator_ attribute. It says the model is not supported. How can I get the Shapley values from the GridSearchCV or something to calculate the Shapley values. One of my columns is categorical, hence the need for preprocessing. Since I have the best_params from the Grid Search I could run the model as an xgboost_regressor model, but it has been a while since doing this without preprocessing.

Inputs:

• Polygon (you can imagine this as a street: long and relatively narrow)
• Line: the line is assumed to lie within the polygon and to run along the full length of the polygon
• Required Area: The area the resulting output sub-polygon must have

Outputs:

• Subpolygon of the input polygon with an area of the required area from input.

Example image of inputs and outputs: input polygon (blue), input line (red), line where the polygon is cut to produce output polygon (green)

The input polygon is cut into two pieces at some point along the given line, with a line that is (if possible) perpendicular to the line.

I hope its clear what I mean - it's already rather difficult to alone describe the problem.

I'm using the shapley geometry library (for python). Polygons are described as a set of points that represent the outer boundary and optionally also sets of point that describe holes inside the polygon. Lines are described as a list of points.

Background of the Problem

I want to explain the outcome of machine learning (ML) models using SHapley Additive exPlanations (SHAP) which is implemented in the shap library of Python. As a parameter of the function shap.Explainer(), I need to pass an ML model (e.g. XGBRegressor()). However, in each iteration of the Leave One Out Cross Validation (LOOCV), the ML model will be different as in each iteration, I am training on a different dataset (1 participant’s data will be different). Also, the model will be different as I am doing feature selection in each iteration.

Then, My Question

In LOOCV, How can I use shap.Explainer() function of shap library to present the performance of a machine learning model? It can be noted that I have checked several tutorials (e.g. this one, this one) and also several questions (e.g. this one) of SO. But I failed to find the answer of the problem.

Thanks for reading!

Update

I know that in LOOCV, the model found in each iteration can be explained by shap.Explainer(). However, as there is 250 participants' data, if I apply shap here for each model, there will be 250 output! Thus, I want to get a single output which will present the performance of the 250 models.

I’m learning JavaScript and I want to implement Gale Shapley algorithm, not exactly that but something similar. A Player orders his choices between available choices and each Choice choose the best ranked players according to the limit of it’s place. Then the rest of players loose their first choices and their second choice become their first one, and the process restart.

I am very new to shapley python package. And I am wondering how should I interpret the shapley value for the Binary Classification problem? Here is what I did so far. Firstly, I used a lightGBM model to fit my data. Something like

I'm using the python shap package to better understand my machine learning model. (From the documentation: "SHAP (SHapley Additive exPlanations) is a game theoretic approach to explain the output of any machine learning model." Below is a small reproducible example of the error I'm getting:

I am following this code https://github.com/BUAA-BDA/FedShapley/tree/master/TensorflowFL and trying to run the file same_OR.py

I also place input file "initial_model_parameters.txt" and data folder "MNIST_data" in same folder