market data of digital asset exchange (binance, kucoin, upbit)

Overview This dataset contains one year of historical cryptocurrency market data collected from three major exchanges: Binance, KuCoin, and Upbit. It was specifically compiled to train, validate, and test the "Feature-Enriched Temporal Graph Neural Networks for Cryptocurrency Arbitrage Detection" (FF-TGN) model. Dataset Details The dataset captures market dynamics and cross-exchange price differences to simulate and detect real-world arbitrage opportunities. It includes essential financial features used in the study, such as price, transaction fees, slippage, liquidity, volume, and network transfer delays, allowing for a realistic backtesting environment that considers time-locks and confirmation delays.