基于ARIMA模型的金融品种走势预测技术  

Financial Variety Trend Prediction Technology Based on ARIMA Model

作　　者：王成国[1] 邓仲元[1] 陈海文[1] 蔡志平[1] 

机构地区：[1]国防科学技术大学计算机学院,湖南长沙410073

出　　处：《计算机技术与发展》2015年第7期11-14,19,共5页Computer Technology and Development

基　　金：国家自然科学基金资助项目(61379145;61170288)

摘　　要：利用金融品种的历史数据开展数据挖掘,有效预测金融品种的走势,为投资者提供决策导向具有广阔的市场前景和应用价值。文中针对金融品种走势预测的应用需求,深入分析金融品种的时间序列特征,总结出其除了包含常见的非线性、非平稳、动态等特征外,还具有高噪音和非正态等特点。基于求和自回归滑动平均模型,建立金融品种走势预测模型,通过实际数据验证了模型的有效性及预测的准确性。自回归滑动平均模型可用于金融品种的动态分析和短期预测。The use of data mining based on historical data of financial products can predict the trend of the financial variety, and it has wide application value and market prospects. According to the application demand of the forecast of the financial variety trend,analyze the characteristics of the time series of the financial variety deeply, and summarize that it not only contains the common characteristics of non -linear, non-stationary and dynamic, but also has the characteristics of high noise and non-normal distribution. Based on autoregressive moving average model, establish the forecasting model of the trend of the financial varieties, and verify the validity of the model and accu- racy of prediction on the basis of the actual data. The autoregressive moving average model can be used to dynamic analysis and make a short-term prediction of financial products.

关 键 词：时间序列 差分自回归滑动平均模型 平稳 金融品种推荐 模型预测 

分 类 号：TP301[自动化与计算机技术—计算机系统结构]