Sybil Attack Address Detection Demo Pipeline

Description

This is a machine learning demo example, convenient for beginners to learn and compete

Specific steps

• Clone this repo. Install the dependencies from the requirements.txt file. If there are any other missing dependencies, please install them yourself.
• Create the following folders
  • ./data/
  • ./saved_model/
  • ./data/raw_data/
  • ./data/features/
• Place all the data files for the competition in the ./data/raw_data/ directory, included:
  • train_dataset.parquet
  • test_dataset.parquet
  • transactions.parquet
  • token_transfers.parquet
  • dex_swaps.parquet
• Run python feature_process.py to generate the feature file at ./data/features/transactions_feature.parquet.
• Run python train.py to train and save model.
• Run python inference.py to generate test dataset prediction result as pred.csv.
• Submit the pred.csv file to the website to view the accuracy of the test set.