COMPARISON OF INVESTMENT PORTFOLIO CONSTRUCTION MODELS

Abstract

Investors all around the world have always been highly interested in portfolio asset allocation problem. Since the 50s of the 20th century, many mathematical models and approaches have been developed, but each of them has its own drawbacks. This task is complicated by the fact that for each individual the decision will be unique: everyone has different preferences and different attitudes towards the risk taken. Financial world is increasingly associated with information technology: the modern securities market cannot operate without automation. A huge number of daily investors’ transactions are processed by IT systems. At the same time, machine learning methods are now at the peak of their popularity and are used in wide variety of fields. The purpose of the article is to compare the results of applying different portfolio construction methods - from classical Markowitz method to decision trees. The initial data are the quotes of some shares listed on the Moscow Exchange.