This repository stores the dataset and the source code used for the experiments presented in the paper
Explainable Ponzi Schemes Detection on Ethereum
The repository is organized as follows:
-
The directory
bytecodescontains the bytecodes of some of the smart contracts used in our analysis. -
The directory
datasetcontains the files to carry on the analysis. In particular, the filesdataset/dataset.csvanddataset/targets.csvcontain the values of features and the relative labels, respectively. -
The directory
notebookcontains the jupyter notebooks used for the paper, more precisely:-
notebook/plot_features.ipynbis used to generates the plots with the features distributions; -
notebook/paper_experiments.ipynbcontains the code to perform the classification experiments described in the paper, and it saves the best classifier as pickle -
notebook/best_classifier_shap_v2.ipynbcontains the code to perform the shap analysis and generate the relative figures taking in input the best classifier obtained through the previous notebook.
-