文献类型：期刊文献

中文题名：基于粗糙集与小波网络集成的股价走势预测研究

英文题名：Research on Integrating Prediction of Stock Price Trend Based on Rough Set and Wavelet Neural Network

作者：任水利[1];雷蕾[2];甘旭升[3];吴亚荣[3]

第一作者：任水利

机构：[1]西京学院;[2]河南财经政法大学工商管理学院;[3]空军工程大学空管领航学院

第一机构：西京学院,西安710123

年份：2017

卷号：37

期号：11

起止页码：2208-2221

中文期刊名：系统科学与数学

外文期刊名：Journal of Systems Science and Mathematical Sciences

收录：北大核心:【北大核心2014】；CSCD:【CSCD2017_2018】；

语种：中文

中文关键词：小波神经网络;粗糙集;属性约简;股票价格;预测

外文关键词：Wavelet neural network, rough set, attribute reduction, stock price,prediction.

摘要：股价走势预测可以为股票投资提供科学依据.为了提高股价走势预测的能力,提出了一种基于粗糙集(RS)与小波网络(WNN)集成的预测方法.它首先利用RS良好的属性约简能力,对股票价格特征量进行降维;然后,采用RS优化WNN的拓扑结构,建立降维特征量基础上的股票价格走势预测模型;最后,由所建模型对股票价格走势进行预测.仿真结果表明,通过引入RS属性约简,很大程度上简化了WNN股格走势模型结构,并改善了模型性能.对上证综指、沪深综指300及澳大利亚股指的预测命中率分别为65.75%、66.37%和65.9%,训练时间为1.7s、1.8s和2.1s,预测结果也优于其它神经网络和WNN模型.从而验证了该方法用于股票价格走势预测的可行性和有效性.
The prediction of stock price trend can provide the scientific basis for stock investment. In order to improve the prediction ability of stock price trend, an integration prediction method based on Rough Set （RS） and Wavelet Neural Network （WNN） is proposed. First, RS attribute reduction is used to reduce the dimensions of feature index for stock price trend, then based on RS attribute reduction the structure of WNN is optimized to establish the prediction model of stock price trend on thebasis of feature index reduction, finally the built model is applied to predict the stock price trend. The simulation results indicate that, by introducing the RS attributes reduction, the structure of WNN model can be simplified to a great extent for stock price trend with improvement of the performance. The direction symmetry of predic- tion corresponding to SSE Composite Index, CSI 300 Index and All Ordinaries Index are 65.75%, 66.37% and 65.93% with training time 1.7s, 1.8s and 2.1s, respectively. The prediction result is better than that of other neural network and WNN models. This verifies the feasibility and effectiveness of the method in the prediction of stock price trend.