Please wait a minute...
高校应用数学学报
Applied Mathematics A Journal of Chinese Universities  2020, Vol. 35 Issue (2): 169-180    DOI:
    
A modified DBSCAN algorithm and its application in finance
HUANG Han-cheng, JIANG Yu
School of Mathematics, Shanghai University of Finance and Economincs, Shanghai 200433
Download:   PDF(410KB)
Export: BibTeX | EndNote (RIS)      

Abstract  This paper presents a modified DBSCAN clustering algorithm with adaptive parameter, and applies it to find potential information clusters of related fund accounts in the stock market.
The algorithm overcome the shortages that parameter ε in the traditional DBSCAN algorithm is oversensitive, and it cannot perform well on multi-densities data sets. Moreover, based on the characteristics
of real data, a new distance is defined to describe the similarity between two funds, which also makes
the modified algorithm better for solving practical problem. Finally, the effectiveness of the modified
algorithm is verified by numerical experiments based on simulated data and real data.

Key words： DBSCAN algorithm      adaptive parameter      mutual fund      information cluster

     
Published: 07 July 2020
CLC:  F830.91  
Service
E-mail this article
Add to my bookshelf
Add to citation manager
E-mail Alert
RSS
Articles by authors
HUANG Han-cheng
JIANG Yu
Cite this article:

HUANG Han-cheng, JIANG Yu. A modified DBSCAN algorithm and its application in finance. Applied Mathematics A Journal of Chinese Universities, 2020, 35(2): 169-180.

URL:

http://www.zjujournals.com/amjcua/     OR     http://www.zjujournals.com/amjcua/Y2020/V35/I2/169


一类改进DBSCAN算法及在金融中的应用

提出了一类具有自适应参数的改进DBSCAN聚类算法, 并应用于发现证券市
场中关联基金账户所组成的信息群落. 算法针对传统算法中半径参数ε敏感度高, 对于
多层密度数据集难以选择全局参数而导致聚类结果差等缺点进行了改进, 此外还基于
实际市场数据特征, 自定义了刻画两个基金间相似程度的综合距离, 使得改进算法能
更好地应用在解决实际问题上. 最后通过基于模拟数据和实际数据的数值实验, 验证
了改进算法的有效性.

关键词： DBSCAN算法,  自适应参数,  公募基金,  信息群落 
[1] WANG Chun-fa, CHEN Rong-da. Option pricing in Markov regime switching Levy models using Fourier-Cosine expansions[J]. Applied Mathematics A Journal of Chinese Universities, 2016, 31(4): 390-404.