Transaction Confirmation Time Estimation in the Bitcoin Blockchain

As Bitcoin is universally recognized as the most popular cryptocurrency, more and more Bitcoin transactions are expected to be populated to the Bitcoin blockchain system. However, transactions cannot be confirmed altogether into the next block due to the limited block capacity. One of the most demanding requirements for users to use Bitcoin is to estimate the confirmation time of a newly submitted transaction. In this paper, we propose two approaches for estimating the confirmation time for a single transaction. The first approach DcyMean makes the estimation based on the historical confirmation time of transactions included in the blockchain. The second approach CTEN is built on neural networks to estimate based on a variety of factors including the transaction itself, block states and mempool states. Finally, we conduct experiments on real Bitcoin blockchain datasets to demonstrate the effectiveness and efficiency of our proposed approaches. Each of our approaches can finish training and estimation within one block interval, demonstrating that our approaches can process real-time cases.