CStock:一种结合新闻与股价的股票走势预测模型

Cstock:A Stock Trend Forecasting Model Combining News and Stock Price

股票是一种高风险、高收益的常见理财产品,为了更好地进行股票投资分析,获得有效的选股方案,文中提出了一种预测股票走势的模型CStock。与现有的股票走势预测模型相比,CStock模型结合新闻和股价走势进行预测,不但利用了股票市场中的交易数据,同时考虑到财经以及政治新闻对于股票市场的影响。CStock模型主要由BiLSTM和CLSTM混合构建,BiLSTM提取股票交易数据的相关特征,CLSTM对新闻的语境特征进行整合和处理,最终通过全连接层输出预测结果。在实验模型中,对股票走势采用分类方法进行实验,得到分类为股票上升的概率和股票下降的概率。实验使用美股数据作为数据集合。通过准确率和收益率进行预测效果评估,实验结果表明,CStock模型在一定程度上能够准确有效地对股票走势进行预测。

Stock is a high-risk,high-yield common financial product.In order to better conduct stock investment analysis and obtain an effective stock selection plan,we propose a model CStock for predicting the stock trend.Compared with the existing stock trend forecasting model,the CStock model combines news and stock price trend to predict.It not only makes use of trading data in the stock market,but also takes into account the influence of financial and political news on the stock market.The CStock model is mainly constructed by mixing BiLSTM and CLSTM.BiLSTM extracts the relevant characteristics of stock trading data,CLSTM integrates and processes the contextual features of news,and finally outputs the predicted results through the fully connected layer.In the experimental model,the stock trend is tested by a classification method,which is classified as the probability of stock rise and the probability of stock decline.The US stock data is used as a data set in the experiment.The prediction results are evaluated by the accuracy rate and the rate of return.The experiment shows that the CStock model can accurately and effectively predict the stock trend to a certain extent.