Problem of constructing portfolio: An application on constructing portfolio with clustering analysis in BIST 100 index

With the increasing globalization rate of the economic system in recent years, the development of financial markets has gained momentum. The growth of exchanges, the increase in public offerings, the emergence of new financial instruments and easy access to them. developments have led to the popularity of stock market investments. However, it is difficult for investors to evaluate the funds in their hands in such a complex market with high uncertainty, where instantaneous changes, rapid and intensive news flows and instant spread to all world markets. However, a number of statistical analysis techniques and algorithms have been developed to help investors make their investment decisions correctly. One of these is clustering analysis which is one of multivariate statistical methods. With clustering analysis, a complex set of data is grouped according to certain variables, making the data more meaningful and it can be preferred in terms of helping to make healthier decisions as a result of the findings. In this study, in order to obtain homogeneous clusters by using clustering analysis on 89 stocks in the Stock Exchange Istanbul 100 Index between 2014- 2018, it was tried to determine the companies that have the best return