Dynamically Controlled Length of Training Data for Sustainable Portfolio Selection

Abstract

Abstract:In a constantly changing market environment, it is a challenge to construct a sustainable portfolio. One cannot use too long or too short training data to select the right portfolio of investments. When analyzing ten types of recent (up to April 2018) extremely high-dimensional time series from automated trading domains, it was discovered that there is no a priori 'optimal' length of training history that would fit all investment tasks. The optimal history length depends of the specificity of the data and varies with time. This statement was also confirmed by the analysis of dozens of multi-dimensional synthetic time series data generated by excitable medium models frequently considered in studies of chaos. An algorithm for determining the optimal length of training history to produce a sustainable portfolio is proposed. Monitoring the size of the learning data can be useful in data mining tasks used in the analysis of sustainability in other research disciplines.

Markowitz’s portfolio theory – optimal length of estimation window for gold and the biggests companies on the Warsaw Stock Exchange

“…Hence, it can be concluded that regardless of the investment objective and the portfolio characteristics studied, the optimal length of the estimation window should be in the range of 144 to 160 observations. Therefore the research hypothesis adopted in the introduction of this paper cannot be rejected, and this is not the same finding as in [Raudys, Raudys, and Pabarskaite, 2018)] when stated that the length of the estimation window depends on the investment purpose. It should be noted that this conclusion was obtained on the basis of daily data and logarithmic rates of return and only for companies with the biggest capitalisation in the WSE and for investment in gold in a given year.…”

“…Despite the literature research, the author could find only one work directly devoted to the issue of the optimal length of the estimation window for portfolio theory. One similar work, alas not connected with stock of big companies and gold but dedicated to automated trading strategies (Raudys, Raudys, and Pabarskaite, 2018), finds that the estimation window length should vary from two to 24 months. The authors also emphasized that the optimal length of the estimation window depends on the investment task and that to the best of their knowledge, this problem has not been considered in the literature.…”

Dantzig Type Optimization Method with Applications to Portfolio Selection

“…Recently, there have been many studies to obtain the strategy for sustainable and effective portfolio selection with various approaches [20][21][22][23]. In this paper, a well-established benchmark (e.g., Dow 30 index), which offers sustainable and stable risk-return profiles with low costs, is used to propose an effective strategy for constructing a portfolio.…”

Immunology-Based Sustainable Portfolio Management

“…In our previous article [11], we examined long time series of multivariate financial data and searched for the 'optimal' length of training data history. We found that the optimal length of history training in everyday trading depends on the specifics of the data and varies over time.…”

“…We found that the optimal length of history training in everyday trading depends on the specifics of the data and varies over time. An algorithm for the calculation of the length of learning history was proposed in [11]. We will also look into day-to-day trading where one or more significant changes in the environment may occur several times a year.…”