<![CDATA[Objective of research to identify the application of LSTM on Ethereum predictions based on blockchain data and information. The study is an experimental study using the Long Short Term Memory (LSTM) method to predict blockchain information on the Ethereum market. The method is a development of Recurrent Neural Network (RNN) and Artificial Neural Network (ANN), and required several precise parameters to produce accurate predictions.The study analyzed several parameters such as the number of neurons in the hidden layer and the most appropriate max epoch to use. The results of the analysis show that using neurons 50 and max epoch 500 are able to predict ethereum prices using blockchain information well, seen from a very small error, namely MAPE of 1.69% with the highest price predictions occurring throughout the middle of 2021 as an effect of changes in the technology system, which is used which previously applied proof of work (mining) to proof of stake (validator) on the current ETH 2.0 technology, as a result there was a decrease in the supply of coins that was currently happening]]>
Copyrights © 2022
