金融时间序列指标判别框架:以特质波动率为例
投稿时间:2015-10-28  修订日期:2016-03-10  点此下载全文
引用本文:
摘要点击次数: 516
全文下载次数: 3
作者单位地址
汤胤 暨南大学管理学院 广东省广州市黄埔大道西601号暨南大学管理学院
毛景慧* 暨南大学管理学院 广州天河区暨南大学学9栋
基金项目:中央高校基本科研业务费专项资金资助项目(15JNLH005);广东省自然科学基金重点项目(2014A030311022)
中文摘要:传统资产组合分析方法往往高度数据敏感,本文提出基于拐点集合判别的TBUD方法。主要思路是分析拐点集合间的关系,并在高维空间进行划分,从而搭建判别模型。论文将分析框架应用在特质波动率等若干指标上,利用实证数据得到结论。应用TBUD判别框架可以发现,特质波动率等指标无法对拐点集合进行清晰划分,因而并不具有预测能力。
中文关键词:特质波动率  支持向量机  贝叶斯判别  趋势预测
 
A Discrimination framework for Financial Time Series Indice based on Inflection Points Set : A Idiosyncratic Volatility Case
Abstract:Traditional Portfolio Analysis is proved to be highly data-dependent. In this paper, we presents a new method which named TBUD to partition the inflection points into collections for time series of the stock price. Analyzing the relation among the collections, this paper build a discrimination framework which applied to Idiosyncratic Volatility, as a case. The result suggests that Idiosyncratic Volatility can not be divided the inflection points set and there for is unable to make an accurate prediction on the future trends of stock price.
keywords:Idiosyncratic Volatility  Support Vector Machine  Bayesian Discrimination  Trend forecasting
查看全文   查看/发表评论   下载pdf阅读器