基于多尺度谱映射的基金投资风格显著特征识别方法  

A Multi-scale Spectral Mapping Based Method to Identify the Significant Features of Funds' Investment Styles

作　　者：苏木亚[1,2] 杨晓光[2] 郭崇慧 Sumuya1,2＇ , Yang Xiaoguang2,Guo Chonghui3(1.School of Economics and Management, Inner Mongolia University, Hohhot 010021; 2.Academy of Mathematics and Systems Science, CAS, Beijing 100190; 3.Facuhy of Management and Economics, Dalian University of Technology, Dalian 11602)

机构地区：[1]内蒙古大学经济管理学院,呼和浩特010021 [2]中国科学院数学与系统科学研究院,北京100190 [3]大连理工大学管理与经济学部,大连116024

出　　处：《管理评论》2018年第3期3-14,共12页Management Review

基　　金：国家自然科学基金项目(61463039;71431008);中国博士后科学基金项目(2015M581192);内蒙古自然科学基金项目(2014BS0706)

摘　　要：风格漂移是证券投资基金常见的现象,辨识投资风格的显著特征对评级和选择证券投资基金有着重要的意义。基于多尺度谱映射的思想,本文提出了一种基金投资风格显著特征识别方法,从逻辑上阐述了该方法的基本原理,用数值模拟展示了该方法相对于传统方法有着明显的优势,并分析了该方法参数选取对识别结果的影响。利用该方法,本文对2012年以来我国24支开放式基金进行了投资风格显著特征的识别,发现这24支基金出现比较严重的风格漂移和趋同现象。Investment style drift is a phenomenon that often occurs in securities investment funds,so identifying the significant features of their investment styles is important for selecting and rating them. Based on spectral mapping idea,we propose a multi-scale spectral mapping based method to identify the significant features of funds＇ investment styles and then logically expound basic principle of the method. Numerical simulation shows that multi-scale spectral mapping is a much more effective method than traditional ones to identify the significant features of funds＇ investment styles. In this paper this method is used to identify the significant features of the key investment styles that 24 Chinese open-end funds have taken on since 2012 and the result shows that there exist obvious convergence and drift in their investment styles.

关 键 词：多尺度谱映射 高斯核参数 投资风格显著特征 

分 类 号：F832.51[经济管理—金融学]