Bitcoin Price Forecasting Using a Long Short-Term Memory Neural Network Model

This study examines the ability to use long-term short-term memory (LSTM) nerve networks to predict bitcoin prices.With increasing instability and investor interest in cryptocurrency, especially bitcoin (BTC), future modeling has attracted considerable attention to financial and academic fields.The study uses historical bitcoin value data received from Yahoo Finance, spreading from January 2019 to June 2025.A sequential LSTM model is manufactured using tensorflow and trained in the last 60 days of closing the price data to predict the concluding value of the next day.The dataset is normalized and divided into training and test sets in a ratio of 80:20.The performance of the model is evaluated using the root mean class error (RMSE), and its predictions are imagined against real prices.Results show that LSTM networks are able to capture temporary price dependence, which provides a promising approach to short-term bitcoin forecasts.This task connects the growing body of literature in the prediction of the financial time chain using deep learning models.