Portfolio selection using complex network in aging society

Portfolio management is an essential problem of financial investment literature. Since Markowitz's portfolio theory introduced, the numerous methods for constructing portfolio set have been proposed in the traditional technology such as the several clustering algorithm and the random matrix theory, while there has NYSE from 01.03 2000 to 12. 31. 2012. To make diverse portfolio sets, we constructed the stock network with winner-takes-all approach. We consider Pearson correlation in order to identify the performance of proposed method and calculated the correlation between the KOSPI and the network- based-portfolio index using in-sample, and out-of-sample return. We find that the correlation value was high enough in overall threshold value in in-sample that can be used as a secondary index. Moreover, we measured portfolio risk using out-of-sample return. We could find that the portfolio's risk much lower than Markowitz's random selected portfolio risk with proposed methodology, which leads to better performance of portfolio in terms of Sharpe ratio. This result implies that we can make sound stock portfolio using complex network.