摘要
在金融二级市场上对数据的分析方法主要是基于统计学和人工建模的方法,本文的提出了使用神经网络的方法分析二级市场金融数据。本文首先提出了在金融二级市场上使用神经网络方法的思路,其次证明了在二级市场使用神经网络进行数据分析的可行性,并且根据金融数据的特点设计出一种适合于处理金融数据的CNN-LSTM网络来处理数据。该网络对比传统的简单的统计方法和一些神经网络方法比如逻辑回归,卷积神经网络(CNN),长短期记忆网络(LSTM)等方法,在对市场价格变化在较短时间内的预测和在较长时间内的预测都有显著的提高,比简单的统计方法提高10%,比其他神经网络提高5%。提出了一种能够较为有效分析金融二级市场数据的方法。
In the secondary market analysis method is mainly based on the statistical and artificial modeling method.We proposed to use neural network to analysis secondary market financial data.First of all,puts forward the idea of using the neural network to analysis the financial secondary market.Secondly,we prove the feasibility of using the neural network to analyze the data in the secondary market,and designs a kind of artificial neural networks which is suitable for dealing with financial data called CNN-LSTM network.Compares to the traditional simple statistical methods and some other neural network methods such as logistic regression,Convolution Neural Network(CNN),Long and Short Term Memory network(LSTM)and other methods,in the market price changes in a relatively short period of time in the forecast and The forecast for a longer period of time has improved significantly,by 10%over simple statistical methods and 5%higher than other neural networks.Proposed a more effective way to analyze the financial secondary market data.
作者
文宇
WEN Yu(Department of Computer Science,ShanghaiJiao Tong University,Shanghai 200240,China)
出处
《电子设计工程》
2018年第17期75-79,84,共6页
Electronic Design Engineering
