Skip to content

Latest commit

 

History

7 Commits

Folders and files

NameName
Last commit message
Last commit date
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

ml-ethereum-fraud

A comparative analysis of dataset characteristics and their impact in ML performance

TL; DR

Read the dataset setup to put the data files required where they belong and run:

./scripts/install-env.sh
source .venv/bin/activate
python src/experiments/showcase.py

Installation

Virtual environment

Run the following command:

chmod +x scripts/install-env.sh
./scripts/install-env.sh

Configuration

The configuration of the project is handled using the params.yaml file. It is recommended to add new information there if required and not hardcoding any values in the code.

WandB

We will use Weights & Biases to track experiments.

If you already have an account: run wandb login.

If you don't have an account contact one of the team members to get one.

Data

Datasets need to be placed in the data directory under the appropriate name. For example, for the dataset in which both ends of the edges belong to the ground truth data, an edge file and a node file need to be placed in the data folder under the names: both_ends_belong_edges.csv and both_ends_belong_nodes.csv.

The exact filename exacted for each dataset can be found inside the src/datasets/variations/ folder.

Running the experiments

All experiments are located inside the src/experiments/ folder. We offer an experiment titled showcase.py that can be used as a guide on how to build new experiments.

About

A comparative analysis of dataset characteristics and their impact in ML performance

Resources

Stars

0 stars

Watchers

1 watching

Forks

Releases

Packages

Contributors

Languages