Maria V. Dobrina, Voronezh State University

Ekaterina A. Kosareva, Voronezh State University

Yana A. Yurova, Voronezh State University

The aim of this work is to search for improved and completely new approaches to the formation of a portfolio of securities that brings maximum profitability. For this, calculations were performed to build the Markowitz investment portfolio, a portfolio based on a probabilistic binary choice model, and a portfolio using the adaptation mechanism under the conditions of the fractal market hypothesis, and a comparative analysis of the obtained calculations was performed. In this work, the authors were able to confirm that the fractal market hypothesis more accurately and correctly describes the processes occurring in financial markets than the effective market hypothesis. To obtain the most optimal and high-quality assessment of the results of the securities portfolio formation, control samples were used. Using the control sample makes it possible to evaluate the predictive properties of the constructed models, to analyze their suitability for applied work on the basis of the data taken, which are unknown at the time the model was formed. The results of a computational experiment demonstrate that the classic Markowitz portfolio in terms of total profitability is significantly inferior to the algorithm portfolio proposed by the authors. In addition, the results of the experiment in the control sample confirm the assumption expressed by the authors in the work that one asset in the financial market can be analyzed as a set of assets according to their investment horizons. The results of experimental calculations also confirmed that the results obtained are acceptable for shares of various issuers and organizations of various fields of activity.