%0 Journal Article
%T 基于CNN-LSTM模型的比特币价格预测<br>Bitcoin Price Prediction Based on CNN-LSTM Model
%A 包娜萍
%A 邢紫豪
%A 夏羽
%J Advances in Applied Mathematics
%P 2956-2966
%@ 2324-8009
%D 2022
%I Hans Publishing
%R 10.12677/AAM.2022.115315
%X 作为具有极度非线性、非平稳性等特征的加密货币，比特币价格预测的研究一直是投资者与研究者的焦点。本文首先选取比特币数据的六个指标，构建LSTM预测模型，发现存在明显滞后性，故进一步构建CNN模型提取数据深层特征，CNN模型拥有更强大的动态捕捉能力，但模型往往存在上下垂直的误差，故最终建立CNN-LSTM混合模型提高预测精度，结果显示无论是模型预测值的总体趋势、局部表现或是滞后性，混合模型都比单一模型更好。<br />
As a cryptocurrency with characteristics such as extreme nonlinearity and non-smoothness, the study of Bitcoin price prediction has been the focus of investors and researchers. This paper first selects six indicators of Bitcoin data and constructs LSTM prediction model, and finds that there is obvious lag, so the CNN model is further constructed to extract deep features of data. CNN model has more powerful dynamic capturing ability, but the model often has up and down vertical error, so the CNN-LSTM is finally established to improve the prediction accuracy, and the results shows that the hybrid model is better than the single model in terms of overall trend, local performance and lagging of the model prediction values.
%K 比特币，长短期记忆神经网络，卷积神经网络<br>Bitcoin
%K LSTM
%K CNN
%U http://www.hanspub.org/journal/PaperInformation.aspx?PaperID=51835