Application of Machine Learning in Prediction of Ethereum-Based Transaction

Abstract

A cryptocurrency payment platform, allows users to transact in cryptocurrencies on a global scale inside a decentralized environment. Ethereum's native token has been successfully migrated to the main net. The platform's digital wallet, Ethereum Wallet, allows users to store and manage their digital assets across several platforms, including computers and mobile devices. Blockchain technology has emerged as a game-changer in aftermath of the success of Bitcoin and other cryptocurrencies. Instead of establishing a new blockchain from scratch as the number of current ones rises, decentralized application developers should focus on finding a solution that best suits their needs and the needs of their decentralized apps. The ownership and transferability of digital financial assets known as 'cryptocurrencies' are guaranteed by decentralized cryptographic technology. The increasing market value and popularity of cryptocurrencies pose a variety of difficulties and concerns for global business and industrial economics. The planned study's main purpose is to validate the correctness of Etherium transactions on the blockchain. The accuracy of the suggested work was compared to that of earlier work in this study. The LSTM-based training strategy has been chosen for the planned study. The training and testing of the Etherium transaction were done with and without considerable record filtering. Both models' accuracy was evaluated to ensure the dependability of the hybrid strategy, which combined an LSTM model with a specific filtering mechanism.