Stock Portfolio Optimization Research Articles

Deep Reinforcement Learning for Stock Portfolio Optimization

Stock portfolio optimization is the process of constant re-distribution of money to a pool of various stocks. In this paper, we will formulate the problem such that we can apply Reinforcement Learning for the task properly. To maintain a realistic assumption about the market, we will incorporate transaction cost and risk factor into the state as well. On top of that, we will apply various state-of-the-art Deep Reinforcement Learning algorithms for comparison. Since the action space is continuous, the realistic formulation were tested under a family of state-of-the-art continuous policy gradients algorithms: Deep Deterministic Policy Gradient (DDPG), Generalized Deterministic Policy Gradient (GDPG) and Proximal Policy Optimization (PPO), where the former two perform much better than the last one. Next, we will present the end-to-end solution for the task with Minimum Variance Portfolio Theory for stock subset selection, and Wavelet Transform for extracting multi-frequency data pattern. Observations and hypothesis were discussed about the results, as well as possible future research directions.1

Optimal stock portfolio diversification under market constraints

The problem of optimal portfolio diversification is considered. Based on mathematical models of the dynamics of the market value formation of a single share and an optimal stock portfolio, the structure of the optimal portfolio is determined. Such models are built in a class of ordinary differential equations. One of the problems of optimal investing is optimizing the expected return of the stock portfolio for the desired level of risk. Another problem is the choice of the stock portfolio with the same expected return, but with a smaller risk. For this purpose, we use a set of acceptable and effective portfolios. This sequence of steps of the algorithm allows consistently solve two optimization problems. The problem of portfolio diversification consists of the problem of determining the moments of time and the necessity to perform such a diversification. In the article, we constructed an algorithm for determining these points of time, based on the solution of an optimal control problem. The application of this algorithm enables to select an optimal risk portfolio at a certain level of its expected profitability. It uses an efficient and acceptable set of investment portfolios.

KRISIS HARGA MINYAK DUNIA DAN KINERJA PORTOFOLIO SAHAM PERUSAHAAN PERKEBUNAN INDONESIA

This study aims to compare the stock portfolio of plantation companies and their performance in the period before and after the crisis of world oil prices. The research method uses the single index model to form a stock portfolio and the index of sharpe, treynor and jensen to measure its portfolio performance . The research uses secondary data of plantation companies listed in the Indonesian Stock Exchange from July 2012 to June 2016. The results showed that there were two plantation company stocks categorized as portfolio candidates in the period before the crisis of world oil prices, namely PT. Astra Agro Lestari Tbk. (AALI) and PT. Tunas Baru Lampung, Tbk. (TBLA), while in the period after the crisis of world oil prices portfolio stocks cannot be established. The stock portfolio of plantation companies in the period before the crisis of wortd oil prices have a higher rate of return and performance index compared to after the crisis of world oil prices. Investors who want to invest in the stock market are advised to make a selection and form an optimal stock portfolio. Investors should observe the movement of stock prices if occur a crisis situation in the national and global levels. DOI : https://doi.org/10.33005/adv.v9i1.2210

A Comprehensive Survey on Portfolio Optimization, Stock Price and Trend Prediction Using Particle Swarm Optimization

Stock market trading has been a subject of interest to investors, academicians, and researchers. Analysis of the inherent non-linear characteristics of stock market data is a challenging task. A large number of learning algorithms are developed to study market behaviours and enhance the prediction accuracy; they have been optimized using swarm and evolutionary computation such as particle swarm optimization (PSO); its global optimization ability with continuous data has been exploited in financial domains. Limitations in the existing approaches and potential future research directions for enhancing PSO-based stock market prediction are discussed. This article aims at balancing the economics and computational intelligence aspects; it also analyzes the superiority of PSO for stock portfolio optimization, stock price and trend prediction, and other related stock market aspects along with implications of PSO.

Comparative Analysis of Mean Variance Efficient Frontier and Resampled Efficient Frontier For Optimal Stock Portfolio Formation

One way to form an optimal stock portfolio is to use the mean-variance efficient frontier (MVEF) method. However, the MVEF method is susceptible to changes in input. An optimal stock portfolio formation method has been developed to overcome this problem, known as the Resampled Efficient Frontier (REF) method. The REF method is a refinement of the MVEF method. REF is done by generating data repeatedly using the Montecarlo simulation. The parameters needed to form a stock portfolio using REF are the average input return μ and the covariance variance matrix Σ. The calculation results show that the weight of the REF portfolio is not negative (there is no short selling strategy). The Sharpe ratio of the REF portfolio is also higher than the MVEF. A larger Sharpe ratio indicates that the REF portfolio has a better performance compared to MVEF. However, empirically based on a case study simulation, the MVEF method experienced a smaller loss compared to the REF method.

A Long Short-Term Memory Approach Towards Stock Selection and Portfolio Optimization

In this paper, we develop a framework for using stock trend prediction outputs, which we generate using long short-term memory (LSTM) deep neural networks, in both stock selection and portfolio optimization. We use LSTM networks to predict the direction of stock movement and a numerical measure of the strength of the stock trend prediction, and use these in stock selection and within the Markowitz mean-variance portfolio optimization framework. Four types of LSTM models are constructed using the Indian SENSEX stock data - individual and ensemble models, each trained using both batch and incremental learning methods. We utilize the accuracy of classification of stock movement direction in shortlisting stocks for the portfolio optimization stage. Diversified and short-selling enabled Markowitz formulations in addition to the standard Markowitz formulation are constructed in the portfolio optimization stage. We also explore the use of a function of the LSTM classification accuracies as a risk measure both in lieu of and in addition to the covariance matrix within the Markowitz framework. Results from each of the above combinations of LSTM construction and portfolio optimization formulation type are benchmarked against the SENSEX and the standard optimal Markowitz portfolios without stock selection. We also analytically derive the conditions under which Markowitz formulations with stock price predictors more accurate than the mean stock price outperform the standard Markowitz formulations. Our work presents a framework that investment analysts can use to incorporate stock trend prediction outputs generated by machine learning techniques in informing their stock selection and optimal portfolio allocation decisions.

Portfolio selection and performance using active and passive strategies (Assessing SRI-KEHATI index in 2013-2018)

This purpose of this study is to examine the simulation results of optimal stock portfolio establishment with active and passive strategy using Tobin’s Qand PBV ratio approach in Sri Kehati Index. We used data from the annual financial reports in 6years2013-2018, and each period, the portfolio was constructed in six fragments; High Tobin’s Q, Medium Tobin’s Q, Low Tobin’s Q, High PBV, Medium PBV and Low PBV. We used Sharpe, Treynor, and Jensen method to measure the portfolio performance and we adjusted for active and passive portfolio strategy and evaluated in each respective period. Finding from this research shows that on passive strategy, the total average of accumulated return is 78.3576%. The low Tobin’s Q portfolio is followed by low PBV, and high PBV has a return value above the total average and above the IDX Composite. While the total average of the accumulated risks in the passive strategy is 23.4193%. Furthermore based on the results of the performance comparison between return and risks from the portfolio and IDX Composite as a market, in general, both are on passive and active strategies. The results also show that the low Tobin’s Q portfolio is consistently able to provide the highest return value, although the low Tobin’s Q portfolio consistently provides the highest risk both on passive and active strategies; thus, there is a consistently between the results of research and the theory of high-risk, high return. Meanwhile the high PBV portfolio, in general, is consistently able to provide a high return, and risk values are consistently at the lowest level compared to other portfolios. The result shows that it has reached the purpose of establishing a stock portfolio

Aplikasi Single Index Model dalam Pembentukan Portofolio Optimal Saham LQ45 pada Bursa Efek Indonesia

Single Index Model is a stock return model that divides the effect on returns into a systematic factor (as measured by yields on the market index) and company-specific factors. The shares used in the formation of the portfolio of the Single Index Model are stocks included in the LQ45 Index in the Indonesia Stock Exchange because this Index has always increased from year to year. However, not all shares in the LQ45 Index can be used, there must be a determination to get the most suitable shares to be used as a portfolio. Therefore, the author uses the Single Index Model Application in the Formation of the LQ45 Optimal Stock Portfolio on the Indonesia Stock Exchange so that investors know how to form an optimal portfolio using a simpler model of the single index model. This study aims to apply the application of a single index model in the formation of a stock portfolio registered in LQ45, determine the proportion of stock investment in the number of lots based on the proportion of stock portfolios, and determine the funds needed for stock investment based on the number of lots. Based on the results of data analysis using the Single Index Model Optimal Portfolio, the results show that out of 45 stocks in LQ45, 13 optimal stocks will be purchased by investors. Thirteen shares and market prices include LPPF Rp.3,400, INTP Rp20,025, LPKR Rp274, INCO Rp3,030, PTPP Rp.2,180, SCMA Rp1,615, TBIG Rp.3,850, AALI Rp.10,350, CTRA Rp1. 070, PWON Rp725, TAXI Rp50, WIKA Rp2,340, and WSKT Rp1,970. The proportion of shares is 9% LPPF, 7% INTP, 5% LPKR, 13% INCO, 19% PTPP, 12% SCMA, 9% TBIG, 7% AALI, 4% CTRA, 6% PWON, 3% TAXI, 4% WIKA , WSKT 2%. The funds to be invested are IDR 1,000,000, the total number of shares to be purchased is 1,271 sheets, or if in the lot lot there are around 12 lots.

Behavioral stock portfolio optimization considering holding periods of B-stocks with short-selling

This study exploits the extra information provided by behavioral stocks (B-stocks) and proposes portfolio selection models that produce profitable portfolios which can outperform traditional investment benchmarks (market index, mutual funds, and exchange-traded funds). Using the embedded holding period (T-day) information of a B-stock, in order to maximize the expected effect (cumulative abnormal return CAR ≥ 1%), this study proposes a buy and hold strategy such that B-stocks are bought when their respective causes are spotted and held until their respective T-days are reached. In addition, in performing the back-tests the actual transaction costs are considered to represent more realistic investments. Short-selling is also considered to provide a more flexible investment such that some B-stocks are shorted when causes are spotted and wait until their respective T-days are reached to exploit the expected effect (CAR≤−1%). The resulting portfolios are then compared to traditional benchmarks. Overall, this study addresses three issues that were not addressed by previous works on B-stocks. These issues are: (1) having portfolios that are profitable even after all transactions costs are considered; (2) identifying short-sell B-stocks for consideration of short-selling in the portfolio selection; and (3) generating portfolios that outperform traditional investment benchmarks.

An intelligence approach for group stock portfolio optimization with a trading mechanism

Optimizing a stock portfolio from a given financial dataset is always a very attractive task, as various factors should be considered. Hence, many methods based on evolutionary algorithms have been developed in the past decades to deal with the portfolio optimization problem. To provide a more flexible stock portfolio, we propose an algorithm to optimize a group stock portfolio by using a grouping genetic algorithm. In accordance with the optimized group stock portfolio, many stock portfolios can be generated and provided to investors. Each chromosome in the genetic algorithm is composed of a grouping part, a stock part and a stock portfolio part. The grouping and stock parts are used to indicate how to divide stocks into groups. The stock portfolio part is used to represent how many stocks should be selected from groups to form a portfolio and what units should be purchased. Four fitness functions are designed to evaluate each individual. Each of them is composed of the group balance, the unit balance, the stock price balance and the portfolio satisfaction. Genetic operations, including crossover, mutation and inversion, are then executed to obtain new offspring to find the best solution. Furthermore, the proposed approach with a trading mechanism is designed to get a more useful group stock portfolio. Experiments on 31 stocks in accordance with four scenarios were conducted to show the merits and effectiveness of the proposed approach.

A novel approach to select the best portfolio considering the preferences of the decision maker

Abstract The challenges of Portfolio Optimization have led to an increasing interest from the multi-objective evolutionary algorithms research community; however, little attention has been paid to the particular preferences of the investor in order to select the most preferred portfolio from a set of mathematically equivalent alternatives in presence of many criteria. The main goal of this work is thus modeling the preferences of the investor in order to find the most satisfactory portfolio from the investor's perspective when many objective functions are considered. Here, the investor's behavior facing risk, the estimations of the portfolios' future returns, and the risk of not attaining those returns are all represented by means of probabilistic confidence intervals. The imperfect knowledge related to the subjectivity of the investor is modeled on the basis of Interval Theory and the outranking method. The proposed approach aggregates the many criteria on the basis of the investor's particular system of preferences producing a selective pressure towards the most preferred portfolio while the investor's cognitive effort in the final selection is reduced. An illustrative example in the context of stock portfolio optimization is provided, where several investors interested in many criteria are simulated. The considered criteria are confidence intervals around the portfolios' expected returns, and indicators from the so-called fundamental and technical analyses. Our approach is compared, using real historical data, with an outstanding multi-objective evolutionary algorithm, MOEA/D, and some well-known benchmarks in Modern Portfolio Theory and Finance Theory, namely, the Mean-Variance approach and the Dow Jones Industrial Average index. The results show an evident superiority of the proposed approach in both the context of the underlying criteria (confidence intervals and financial indicators) and the context of the actual returns. Thus, we conclude that the proposed approach was able to find satisfactory portfolios in the context of the experiments.

An Effective Approach for the Diverse Group Stock Portfolio Optimization Using Grouping Genetic Algorithm

Finding useful portfolios that could be a portfolio of trading strategy or a stock portfolio from financial datasets is always an attractive research topic due to the nature of financial markets. Because investors always want an approach that can continually provide various portfolios, the issue of group stock portfolio optimization (GSPO) has been raised and the algorithms to obtain a group stock portfolio (GSP) have also been described in the past. A GSP divides the whole set of stocks into several stock groups, and the stocks in a group are exchangeable in investment. Thus, when investors are not satisfied with a suggested stock, they can select another stock from the same group to replace the original one. However, the industry diversity of stocks within a group is not regarded in the existing literatures. In this paper, an algorithm for dealing with the diverse group stock portfolio optimization (DGSPO) is proposed to obtain a diverse group stock portfolio (DGSP). The proposed algorithm is based on the group genetic algorithm with the chromosome representation and the fitness function designed for the purpose of finding a good DGSP. Especially, a factor called group diversity is designed to diversify stocks from different industries and is considered in the fitness evaluation. Another factor considers cash dividend is also applied to keep companies with good quality in the portfolio for increasing the profit of a DGSP. Two real financial datasets are used in the experiments to verify the effectiveness of the proposed approach.

МОДЕЛИРОВАНИЕ ДОХОДНОСТИ АКЦИЙ КОМПАНИЙ РАЗЛИЧНОЙ СТРАНОВОЙ И ОТРАСЛЕВОЙ ПРИНАДЛЕЖНОСТИ

The article considers a method of modeling of a stock portfolio’s returns, involving separation of industrial and systematic risk at the global and national levels. Various types of risk get distinguished by introduction of industrial and country stock indices into the structure of the pair copula. The proposed method allows combining advantages of the traditional CAPM and pair-copula structures. In the frameworks of solving the problem of a stock portfolio optimization, two methods of formation of the structure made of pair copulas have been tested: the traditional one that ensures the maximum of the sum of Kendall’s correlation ratios between return on assets, and the method of structuring based on industrial and country principle featuring introduction of stock indices into the model. The minimum of risk and the maximum of the Sharpe ratio were considered as optimality criteria of the stock portfolio. The model, structured by the industrial and country principles, showed a greater capital gain for the both indices. The proposed method of accounting the industrial and systematic risk can be useful for professional managers when forming an investment portfolio and improving its returns.

Optimal Stock Portfolio Issuers of Building Construction Registered in LQ45 Based on the Markowitz Approach

Optimal Stock Portfolio Issuers of Building Construction Registered in LQ45 Based on the Markowitz Approach

Creating Dynamic Pretrade Models: Beyond the Black Box

We provide a framework for investment managers to create dynamic pretrade models. The approach helps market participants shed light on vendor black-box models that often do not provide any transparency into the model’s functional form or working mechanics. In addition, this allows portfolio managers to create consensus estimates based on their own expectations, such as forecasted liquidity and volatility, and to incorporate firm proprietary alpha estimates into the solution. These techniques allow managers to reduce overdependency on any one black-box model, incorporate costs into the stock selection and portfolio optimization phase of the investment cycle, and perform “what-if” and sensitivity analyses without the risk of information leakage to any outside party or vendor.

An Optimal Stock Market Portfolio Proportion Model Using Genetic Algorithm

To reduce the amount of loss due to investment risk, an investor or stockbroker usually forms an optimal stock portfolio. This technique is done to get the maximum return of investment on shares to be purchased. However, in forming a stock portfolio required a fairly complex calculations and certain skills. This work aims to provide an alternative solution in the problem of forming the optimal and efficient stock portfolio composition by designing a system that can help decision making of investors or stockbrokers in preparing stock portfolio in accordance with the policy and risk investment. In this work, determination of optimal stock portfolio composition is constructed by using Genetic Algorithm. The data used in this work are the 4 selected stocks listed on the LQ45 index in 2017. Meanwhile, the calculation of profit and loss rate utilizes a single index model theory. The efficiency of the algorithm has been examined against the population size and crossover and mutation probabilities. The experimental results show that the proposed algorithm can be used as one of solutions to select the optimal stock portfolio.

Optimal Portfolio Allocation based on two Novel Risk Measures and Genetic Algorithm

The problem of optimal portfolio selection has attracted a great attention in the finance and optimization field. The future stock price should be predicted in an acceptable precision, and a suitable model and criterion for risk and the expected return of the stock portfolio should be proposed in order to solve the optimization problem. In this paper, two new criterions for the risk of stock price prediction has been presented, of which the first one is based on the interval predictions which vary with time and proportional to the uncertainty of stock price data, while the second one is a constant risk term that is proportional to the prediction error variances of the neural networks. A novel cost function has been presented to simultaneously consider the expected returns and risks. Genetic algorithm has been used to solve this optimization problem. Finally, 18 shares of the Tehran Stock Exchange have been considered to evaluate the performance of the proposed risk criterions. Two proposed risk criteria, by the conditional value at risk (CVaR) associated with the same stock. The problem of stock portfolio optimization has been solved for all three situations, and the PI-based risk criteria yielded a better return

Stock Portfolio-Optimization Model by Mean-Semi-Variance Approach Using of Firefly Algorithm and Imperialist Competitive Algorithm

Selecting approaches with appropriate accuracy and suitable speed for the purpose of making decision is one of the managers’ challenges. Also investing decision is one of the main decisions of managers and it can be referred to securities transaction in financial markets which is one of the investments approaches. When some assets and barriers of real world have been considered, optimization of stock basket can’t be solved easily, therefore, Meta-Heuristic approach is considered. In this regard, the main goal of this paper is to solve stock portfolio constrained optimization problem by using of Firefly Algorithm (FA) and Imperialist Competitive Algorithm (ICA). In order to do so, daily information of 25 accepted stocks in period of 2010-2016, in Tehran stock market have been used. Results show that Firefly Algorithm (FA) and Imperialist Competitive Algorithm (ICA) showed successful function in constrained optimization of stock portfolio and has acceptable accuracy in finding optimal answers in whole risk and returns levels. Also, the results of comparison of Cardinality Constrained Mean – Variance (CCMV) and Cardinality Constrained Mean – Semi -Variance (CCMSV) portfolios two using of Firefly Algorithm (FA) and Imperialist Competitive Algorithm (ICA) and considering to the findings of two criteria for assessing accuracy in stock basket optimization simultaneously show that Imperialist Competitive Algorithm (ICA) has high speed and accuracy for solving stock basket optimization and could have desirable interaction with real barriers of market. Moreover, there is high accuracy optimization of Cardinality Constrained Mean – Semi -Variance (CCMSV) compared to Cardinality Constrained Mean – Variance (CCMV).

Behavioral Stock Portfolio Optimization through Short-Selling

Behavioral Stock Portfolio Optimization through Short-Selling

Application of Interval Valued Fuzzy Linear Programming for Stock Portfolio Optimization

In this paper, based on existing results, decision making about portfolio investment schemes is discussed, ordering method of fuzzy numbers of interval value is shown, corresponding auxiliary models are established and solutions are provided with theories of fuzzy mathematics, optimization theory and numerical calculation, etc. Then it applies software programming to solve the portfolio investment situation between investors in savings and four securities according to the established models. The result shows that investors can choose the risk coefficient that they can bear to reach the maximum value of expected returns. The greater the risk coefficient, the greater the income, the smaller the risk coefficient and the smaller the income. Investors can determine their own portfolio strategy according to their own conditions in order to meet their own interests.