Prediction And Analysis Of Bitcoin Price Based On Lstm Neural Network Under The Framework Of Deep Learning

In recent years,with the development of computer technology,metaverse,Web3.0and other emerging fields gradually come into public view.To put it simply,they are the decentralized Internet based on blockchain technology,which quickly lands in distributed storage,open finance,digital currency and other fields.As the first application of blockchain technology,Bitcoin has been widely studied and concerned since its birth.With its price rising,many scholars began to try to analyze and forecast its price.In the early days,they used a time-series model to predict the price of bitcoin,which is good for linear data but not good for non-linear data like bitcoin.With the rapid development of deep learning,many scholars begin to use RNN and LSTM neural networks to predict the price of bitcoin.RNN,LSTM and other neural networks have memory for sequences,can retain useful information,and can fit nonlinear data well,so some achievements have been made.This paper summarizes four factors that may affect the price of bitcoin,and it is these factors that cause the game between buyers and sellers and then lead to the fluctuation and nonlinearity of the price of bitcoin.It respectively introduces the theoretical knowledge of the time series ARIMA model and LSTM neural network model,and collects the real data of the bitcoin market,conducts modeling and analysis on it,and predicts its price.Then the prediction results of different models are compared by root mean square error.Due to the high requirement of ARIMA model on sample data,it can not fully explore the nonlinear components in time series,so the prediction effect is poor.LSTM neural network can selectively memorize useful information in time series and transmit the previous information,so it is more suitable for nonlinear data prediction and fitting.The results show that the multivariate LSTM neural network model has the best prediction effect,which is better than the univariate LSTM neural network model and better than the ARIMA model.It shows that the LSTM neural network is a better method to predict the price of bitcoin.