资源描述:
《基于最小生成树的多因子选股》由会员上传分享,免费在线阅读,更多相关内容在学术论文-天天文库。
1、分类号:密级:UDC:编号:201421108002河北工业大学硕士学位论文基于最小生成树的多因子选股论文作者:郑向坤学生类别:全日制学科门类:理学硕士学科专业:统计学指导教师:陈爽职称:副教授DissertationSubmittedtoHebeiUniversityofTechnologyforTheMasterDegreeofScienceinAppliedMathematicsMulti-factormodelbasedontheminimalspanningtreebyXiangkunZhengSupervi
2、sor:A.P.ChenShuangDecember2016摘要近年来,人们对量化投资有了深入的了解。其中,多因子选股模型是一种特别重要的模型,它的优势在于能够结合很多信息,得出人们理想中的选股结果。在经济物理学中,股票网络是曼泰尼亚引入的,是运用最小生成树的方法来调查股票之间的联系。为了研究股票之间的相互联系,本文使用最小生成树的方法探索股票网络的结构。本文首先通过对股票投资者访问,探索投资者选取股票倾向于选取哪些因子作为参考标准,也就是候选因子的选取。然后,本文运用多因子选股的方法,经过有效因子的检验,剔除有效且冗
3、余的因子。最后,本文采用三种方法,即Kruskal算法、Prim算法、权矩阵方法,求解股票网络的最小生成树。根据分组技术对股票的因子进行建模,模型求解,找出股票的相异度。关键字:量化投资多因子选股模型最小生成树Kruskal算法Prim算法权矩阵IABSTRACTInrecentyears,quantitativeinvestmenthadbeenstudiedwidely.Amongthosegeneralizations,multi-factormodelhasbecomeanincreasinglypopular
4、model,theadvantageofthemodelisthecombinationoftheinformation,wecanobtaintheidealresultthroughthis.Intheeconophysics,thestocknetworkwasfirstmentionedbyMantegna,wheretheminimumspanningtreewasusedtoinvestigatetheconnectionofthestock.Inordertostudytheconnectionofthe
5、stock,weusetheminimumspanningtreetoresearchthecontructionofthestocknetwork.Inthispaper,wefirstlyseekthechoiceoftheinvestorswhichtendtoselectsomefactorsasareferencefactorbyvisitingthestockinvestors,thatistosaythechoiceofthecandidatefactors.Then,wecanusethemulti-f
6、actormethodtotakeouttheeffectivefactorsthroughtheeffectivefactorsofinspection.Furthermore,weemploythreemeathodstosolvetheminimumspanningtreeofthestocknetwork,includingkruskalalgorithm,primalgorithm,weightmatrix.Accordingtothegrouptechnology,wecanmodelthefactorof
7、thestock,seekthesolutionofthemodelandfindoutthestockofthedissimilarity.KEYWORDS:quantitativeinvestmentmulti-factormodelminimumspanningtreeKruskalalgorithmPrimalgorithmweightmatrixII目录摘要.............................................................................
8、....................................................IIIABSTRACT.......................................................................................................