Özet:

Portfolio selection is a process of choice that is important in the field of economy and finance. The classical modern portfolio theory is a model focused on portfolio return and risk in the light of the default historical data in which portfolio choice problem is normally distributed. On the other hand, the previous period revenue series of shares are often not normally dispersed in real life, and it is meaningful to add the model of confusion and pressure. With high-degree moments and portfolio optimization, the entropic function is added to the model, which provides natural diversity to prevent accumulation in certain stocks and to include future uncertainty in the model. In the study, the model, which has become np-hard with the addition of quantity limit that restricts the number of shares that can be found in the portfolio, has been solved with the optimization of the particle bulk. From the stock assets found in the sample data set, models for different scenarios have been set up and the effectiveness in the diversification of the entropic function, which is recommended for the selection process, has been discussed.

Anahtar Kelimeler:

Özet:

Portfolio selection is an important selection process in economy and finance. The classic modern portfolio theory is a model that focuses on portfolio return and risk in the light of historical data that is assumed to be normally distributed in portfolio selection problem. However, the past return series of stocks are not normally distributed frequently in real life, and it is meaningful to add skewness and kurtosis to the portfolio model. The entropy function that provides the natural diversity is included in the model in order to add future uncertainty in the model and prevent the accumulation of certain stocks encountered in portfolio selection based on higher order moments. In the study, the model, which has become a np-hard problem with the addition of cardinality constraint limiting the number of stocks that can be found in the portfolio, has been solved by particle swarm optimization. From the assets in the sample dataset, models were set for different scenarios and the effectiveness of the proposed entropy function for selection process, in diversification was discussed.