|国家预印本平台
首页|非参数回归模型在金融时间序列上的应用

非参数回归模型在金融时间序列上的应用

pplication for Nonparametric Regression Models in the Financial Time Series

中文摘要英文摘要

本文旨在运用非参数回归模型来解决金融上的实际问题;对1998~2009年的上证综指的收益率数据进行了简单的统计分析,说明利用非参数回归方法来研究股市收益率的问题较合理。然后,分别运用核回归和LOWESS预测了2010年上证综指的收盘价和收益率,结果显示预测效果较佳,核回归拟合曲线较为平滑,但仍存在着边界效应。采用了平均相对误差和RMSE分别检测两种模型的预测效果,结果显示核回归的预测效果优于LOWESS。

Ihis paper aims at using the nonparametric regression models to solve the practical financial issue. In the part of empirical analysis, the paper does some basic statistical analysis for the return ratio of Shanghai stock market from 1998 to 2009, the result shows applying nonparametric regression in studying the return ratio of stock market is more reasonable. Next, the paper uses the kernel regression and LOWESS to predict the closing price and yield rates of Shanghai stock market in 2010 separately, and the effect of the two forcast curve is ideal. The average relative error and RMSE are used to test the prediction effects of the two models separately. Seeing from these two index, the prediction effect of the kernel regression is better than that of LOWESS.

刘琼荪、朱云霓

财政、金融

金融时间序列非参数回归LOWESS核回归V

Financial Time SeriesNonparametric RegressionLOWESSKernel RegressionV

刘琼荪,朱云霓.非参数回归模型在金融时间序列上的应用[EB/OL].(2014-12-15)[2025-08-18].http://www.paper.edu.cn/releasepaper/content/201412-421.点此复制

评论