Open AccessROBUST PORTFOLIO SELECTION WITH CLUSTERING BASED ON BUSINESS SECTOR OF STOCKS
Author(s) -
La Gubu,
Dedi Rosadi,
Abdurakhman Abdurakhman
Publication year - 2021
Publication title -
media statistika
Language(s) - English
Resource type - Journals
ISSN - 2477-0647
DOI - 10.14710/medstat.14.1.33-43
Subject(s) - portfolio , cluster analysis , outlier , sharpe ratio , portfolio optimization , econometrics , covariance matrix , computer science , selection (genetic algorithm) , efficient frontier , stock (firearms) , statistics , mathematics , economics , financial economics , artificial intelligence , engineering , mechanical engineering
In recent years there have been numerous studies on portfolio selection using cluster analysis in conjunction with Markowitz model which used mean vectors and covariance matrix that are estimated from a highly volatile data. This study presents a more robust way of portfolio selection where stocks are grouped into clusters based on business sector of stocks. A representative from each cluster is selected from each cluster using Sharpe ratio to construct a portfolio and then optimized using robust FCMD and S-estimation. Calculation Sharpe ratio showed that this method works efficiently on large number of data while also robust against outlier in comparison to k-mean clustering. Implementation of this method on stocks listed on the Indonesia Stock Exchange, which included in the LQ-45 indexed for the period of August 2017 to July 2018 showed that portfolio performance obtained using clustering base on business sector of stocks combine with robust FMCD estimation is outperformed the other possible combination of the methods.