基于小波神经网络与ARIMA组合模型在股票预测中的应用

杨 进; 陈 亮

基于小波神经网络与ARIMA组合模型在股票预测中的应用

引用本文：杨进,陈亮.基于小波神经网络与ARIMA组合模型在股票预测中的应用[J].经济数学,2018,(2):62-67

摘要点击次数: 1081

全文下载次数: 0

作者	单位
杨进,陈亮	（上海理工大学理学院上海 200093）

中文摘要:为了实现对股票价格变化的短期预测，提出了一种基于小波神经网络（WNN）与自回归积分滑动平均模型（ARIMA）的组合预测模型.将股票的收盘价序列数据划分为线性以及非线性（误差项）两个部分，分别利用统计学中ARIMA模型和小波神经网络分别对两部分数据进行预测并得到结果，将两部分结果组合相加合成为整个股票价格的预测结果.实验结果表明该组合模型在预测精度方面有提高，是一种比较有效的预测模型.

中文关键词:应用数学；组合预测股票价格 ARIMA模型；小波神经网络

The Application in Stock Prediction of Combination Forecast Model of Wavelet Neural Network and ARIMA

Abstract:To realize the prediction capacity of short-term stock price, an integrated prediction model is proposed based on Wavelet Neural Network (WNN) and ARIMA. The stock price data is firstly divided into linear parts and nonlinear parts. On this basis, ARIMA model is used to predict the linear parts and get the first result. WNN model is used to predict the nonlinear parts and get another result. Finally, the final prediction result of the stock price is obtained by adding the two partial results. The final prediction result is better than those obtained by other neural networks, which verifies the feasibility and effectiveness of the proposed model of predicting stock price.

keywords:applied mathematics combination forecast for stock price ARIMA model wavelet neural network

查看全文 查看/发表评论 下载pdf阅读器