scholarly journals Can Machine Learning-Based Portfolios Outperform Traditional Risk-Based Portfolios? The Need to Account for Covariance Misspecification

Risks ◽  
2019 ◽  
Vol 7 (3) ◽  
pp. 74 ◽  
Author(s):  
Prayut Jain ◽  
Shashi Jain
Keyword(s):  
Machine Learning ◽  
Predictive Ability ◽  
Forecast Model ◽  
Minimum Variance ◽  
Portfolio Performance ◽  
Risk Parity ◽  
Out Of Sample ◽  
Weighted Portfolio ◽  
The Impact ◽  
Allocation Methods

The Hierarchical risk parity (HRP) approach of portfolio allocation, introduced by Lopez de Prado (2016), applies graph theory and machine learning to build a diversified portfolio. Like the traditional risk-based allocation methods, HRP is also a function of the estimate of the covariance matrix, however, it does not require its invertibility. In this paper, we first study the impact of covariance misspecification on the performance of the different allocation methods. Next, we study under an appropriate covariance forecast model whether the machine learning based HRP outperforms the traditional risk-based portfolios. For our analysis, we use the test for superior predictive ability on out-of-sample portfolio performance, to determine whether the observed excess performance is significant or if it occurred by chance. We find that when the covariance estimates are crude, inverse volatility weighted portfolios are more robust, followed by the machine learning-based portfolios. Minimum variance and maximum diversification are most sensitive to covariance misspecification. HRP follows the middle ground; it is less sensitive to covariance misspecification when compared with minimum variance or maximum diversification portfolio, while it is not as robust as the inverse volatility weighed portfolio. We also study the impact of the different rebalancing horizon and how the portfolios compare against a market-capitalization weighted portfolio.

Forecasting US movies box office performances in Turkey using machine learning algorithms

2020 ◽  
Vol 39 (5) ◽  
pp. 6579-6590
Author(s):  
Sandy Çağlıyor ◽  
Başar Öztayşi ◽  
Selime Sezgin
Keyword(s):  
Machine Learning ◽  
Global Economy ◽  
Learning Algorithms ◽  
Forecast Model ◽  
Machine Learning Algorithms ◽  
Gradient Boosting ◽  
High Stakes ◽  
Box Office ◽  
Industry Forecast ◽  
The Impact

The motion picture industry is one of the largest industries worldwide and has significant importance in the global economy. Considering the high stakes and high risks in the industry, forecast models and decision support systems are gaining importance. Several attempts have been made to estimate the theatrical performance of a movie before or at the early stages of its release. Nevertheless, these models are mostly used for predicting domestic performances and the industry still struggles to predict box office performances in overseas markets. In this study, the aim is to design a forecast model using different machine learning algorithms to estimate the theatrical success of US movies in Turkey. From various sources, a dataset of 1559 movies is constructed. Firstly, independent variables are grouped as pre-release, distributor type, and international distribution based on their characteristic. The number of attendances is discretized into three classes. Four popular machine learning algorithms, artificial neural networks, decision tree regression and gradient boosting tree and random forest are employed, and the impact of each group is observed by compared by the performance models. Then the number of target classes is increased into five and eight and results are compared with the previously developed models in the literature.

On portfolio optimization

2016 ◽  
Vol 17 (3) ◽  
pp. 295-309 ◽  
Author(s):  
Theo Berger ◽  
Christian Fieberg
Keyword(s):  
Minimum Variance ◽  
Factor Models ◽  
Scale Factor ◽  
Covariance Matrices ◽  
Portfolio Performance ◽  
Content Type ◽  
Portfolio Decisions ◽  
Multi Scale ◽  
Out Of Sample ◽  
Factor Returns

Purpose The purpose of this paper is to show how investors can incorporate the multi-scale nature of asset and factor returns into their portfolio decisions and to evaluate the out-of-sample performance of such strategies. Design/methodology/approach The authors decompose daily return series of common risk factors and of all stocks listed in the Dow Jones Industrial Index (DJI) from 2000 to 2015 into different time scales to separate short-term noise from long-run trends. Then, the authors apply various (multi-scale) factor models to determine variance-covariance matrices which are used for minimum variance portfolio selection. Finally, the portfolios are evaluated by their out-of-sample performance. Findings The authors find that portfolios which are constructed on variance-covariance matrices stemming from multi-scale factor models outperform portfolio allocations which do not take the multi-scale nature of asset and factor returns into account. Practical implications The results of this paper provide evidence that accounting for the multi-scale nature of return distributions in portfolio decisions might be a promising approach from a portfolio performance perspective. Originality/value The authors demonstrate how investors can incorporate the multi-scale nature of returns into their portfolio decisions by applying wavelet filter techniques.

scholarly journals The impact of data-complexity and team characteristics on performance in the classification model

2022 ◽  
Vol 9 (1) ◽  
pp. 0-0
Keyword(s):  
Machine Learning ◽  
Binary Classification ◽  
Class Imbalance ◽  
Predictive Ability ◽  
Predictive Performance ◽  
Classification Model ◽  
Data Complexity ◽  
High Performing ◽  
Machine Learning Model ◽  
The Impact

This article investigates the impact of data-complexity and team-specific characteristics on machine learning competition scores. Data from five real-world binary classification competitions hosted on Kaggle.com were analyzed. The data-complexity characteristics were measured in four aspects including standard measures, sparsity measures, class imbalance measures, and feature-based measures. The results showed that the higher the level of the data-complexity characteristics was, the lower the predictive ability of the machine learning model was as well. Our empirical evidence revealed that the imbalance ratio of the target variable was the most important factor and exhibited a nonlinear relationship with the model’s predictive abilities. The imbalance ratio adversely affected the predictive performance when it reached a certain level. However, mixed results were found for the impact of team-specific characteristics measured by team size, team expertise, and the number of submissions on team performance. For high-performing teams, these factors had no impact on team score.

scholarly journals Técnicas Quantitativas de Otimização de Carteiras Aplicadas ao Mercado de Ações Brasileiro

2012 ◽  
Vol 10 (3) ◽  
pp. 369
Author(s):  
André Alves Portela Santos ◽  
Cristina Tessari
Keyword(s):  
Covariance Matrix ◽  
Minimum Variance ◽  
Short Selling ◽  
Optimization Techniques ◽  
Market Portfolio ◽  
Sample Covariance ◽  
Out Of Sample ◽  
Weighted Portfolio ◽  
Performance Statistics ◽  
Mean Variance

In this paper we assess the out-of-sample performance of two alternative quantitative portfolio optimization techniques - mean-variance and minimum variance optimization – and compare their performance with respect to a naive 1/N (or equally-weighted) portfolio and also to the market portfolio given by the Ibovespa. We focus on short selling-constrained portfolios and consider alternative estimators for the covariance matrices: sample covariance matrix, RiskMetrics, and three covariance estimators proposed by Ledoit and Wolf (2003), Ledoit and Wolf (2004a) and Ledoit and Wolf (2004b). Taking into account alternative portfolio re-balancing frequencies, we compute out-of-sample performance statistics which indicate that the quantitative approaches delivered improved results in terms of lower portfolio volatility and better risk-adjusted returns. Moreover, the use of more sophisticated estimators for the covariance matrix generated optimal portfolios with lower turnover over time.

scholarly journals Factor-Based Optimization of a Fundamentally-Weighted Portfolio in the Illiquid and Undeveloped Stock Market

2020 ◽  
Vol 13 (12) ◽  
pp. 302
Author(s):  
Davor Zoričić ◽  
Denis Dolinar ◽  
Zrinka Lovretin Golubić
Keyword(s):  
Risk Factors ◽  
Stock Market ◽  
Principal Component ◽  
Minimum Variance ◽  
Portfolio Risk ◽  
Long Run ◽  
Out Of Sample ◽  
Weighted Portfolio ◽  
Sample Testing ◽  
Underlying Risk

In this paper, the possibility of using fundamental weighting as a tool to intentionally tilt a portfolio toward specific and unobservable risk factors in the illiquid and undeveloped Croatian stock market is explored. Thus far, fundamental-weighting has been shown to be able to outperform the cap-weighted index in such environments but no attempt regarding control for implicit factor exposure of such portfolios has been reported. Therefore, in this study principal component analysis is performed to capture the underlying risk factors of the fundamentally-weighted portfolio in order to optimize the portfolio’s performance by minimizing its volatility. Previous attempts focusing purely on portfolio risk reduction by estimating minimum variance portfolios failed both from an in-sample and out-of-sample perspective. Results in this study are based on 22 in-sample and out-of-sample tests in the period from March 2009 till March 2020. On the in-sample estimation basis, the proposed approach significantly improves the portfolio’s performance and, if restrictions to weights are imposed, it can outperform the cap-weighted benchmark. However, out-of-sample testing yielded poor results both in terms of risk and return. Such results are in contrast to findings for the developed markets but corroborate the claim that a broad investment base is needed for successful risk exposure in the long run.

A Cross-Sectional Machine Learning Approach for Hedge Fund Return Prediction and Selection

2020 ◽  
Author(s):  
Wenbo Wu ◽  
Jiaqi Chen ◽  
Zhibin (Ben) Yang ◽  
Michael L. Tindall
Keyword(s):  
Machine Learning ◽  
Hedge Fund ◽  
Forecast Model ◽  
Specific Information ◽  
Forecast Method ◽  
Learning Methods ◽  
Time Series Regression ◽  
Cross Sectional ◽  
Machine Learning Methods ◽  
Out Of Sample

We apply four machine learning methods to cross-sectional return prediction for hedge fund selection. We equip the forecast model with a set of idiosyncratic features, which are derived from historical returns of a hedge fund and capture a variety of fund-specific information. Evaluating the out-of-sample performance, we find that our forecast method significantly outperforms the four styled Hedge Fund Research indices in almost all situations. Among the four machine learning methods, we find that deep neural network appears to be overall most effective. Investigating the source of methodological advantage of our method using a case study, we find that cross-sectional forecast outperforms forecast based on time series regression in most cases. Advanced modeling capabilities of machine learning further enhance these advantages. We find that the return-based features lead to higher returns than the benchmark of a set of macroderivative features, and our forecast method yields best performance when the two sets of features are combined. This paper was accepted by David Simchi‐Levi, finance.

scholarly journals The Impact of Advertising and R&D on Bankruptcy Survival: A Double-Edged Sword

2020 ◽  
Vol 84 (5) ◽  
pp. 22-40
Author(s):  
Niket Jindal
Keyword(s):  
Machine Learning ◽  
Research And Development ◽  
Cash Flow ◽  
The Other ◽  
Cumulative Effects ◽  
Future Cash Flow ◽  
Low Level ◽  
Out Of Sample ◽  
The One ◽  
The Impact

Advertising and research and development (R&D) benefit firms by increasing sales and shareholder value. However, when a firm is in bankruptcy, the cumulative effects of its past advertising and R&D can be a double-edged sword. On the one hand, they increase the firm’s expected future cash flow, which increases the likelihood that the bankruptcy court will decide the firm can survive. On the other hand, they increase the liquidation value of the firm’s assets, which decreases the likelihood that the bankruptcy court will decide that the firm can survive. The author argues that the ability of advertising and R&D to either increase or decrease bankruptcy survival is contingent on the influence that the firm’s suppliers have, relative to other creditors, on the bankruptcy court’s decision. Advertising and R&D increase (decrease) bankruptcy survival when suppliers have a high (low) level of influence. Empirical analyses, conducted on 1,504 bankruptcies, show that advertising (R&D) increases bankruptcy survival when at least 35%−38% (18%−21%) of the bankrupt firm’s debt has been borrowed from suppliers, whereas it decreases bankruptcy survival below this point. Out-of-sample machine learning validation shows that the ability to predict whether a bankrupt customer will survive is substantially improved by considering the firm’s advertising and R&D.

scholarly journals A Combined Control Systems and Machine Learning Approach to Forecasting Iceberg Flux off Newfoundland

2021 ◽  
Vol 13 (14) ◽  
pp. 7705
Author(s):  
Jennifer B. Ross ◽  
Grant R. Bigg ◽  
Yifan Zhao ◽  
Edward Hanna
Keyword(s):  
Machine Learning ◽  
Control Systems ◽  
Late Summer ◽  
Forecast Model ◽  
Seasonal Forecast ◽  
Rate Of Change ◽  
Learning Tools ◽  
Systems Model ◽  
The North ◽  
The Impact

Icebergs have long been a threat to shipping in the NW Atlantic and the iceberg season of February to late summer is monitored closely by the International Ice Patrol. However, reliable predictions of the severity of a season several months in advance would be useful for planning monitoring strategies and also for shipping companies in designing optimal routes across the North Atlantic for specific years. A seasonal forecast model of the build-up of seasonal iceberg numbers has recently become available, beginning to enable this longer-term planning of marine operations. Here we discuss extension of this control systems model to include more recent years within the trial ensemble sample set and also increasing the number of measures of the iceberg season that are considered within the forecast. These new measures include the seasonal iceberg total, the rate of change of the seasonal increase, the number of peaks in iceberg numbers experienced within a given season, and the timing of the peak(s). They are predicted by a range of machine learning tools. The skill levels of the new measures are tested, as is the impact of the extensions to the existing seasonal forecast model. We present a forecast for the 2021 iceberg season, predicting a medium iceberg year.

scholarly journals “Go Wild for a While!”: A New Test for Forecast Evaluation in Nested Models

Mathematics ◽  
2021 ◽  
Vol 9 (18) ◽  
pp. 2254
Author(s):  
Pablo Pincheira ◽  
Nicolás Hardy ◽  
Felipe Muñoz
Keyword(s):  
Predictive Ability ◽  
Random Variable ◽  
Simple Modification ◽  
Nested Models ◽  
Asymptotically Normal ◽  
Estimation Uncertainty ◽  
Normal Test ◽  
Out Of Sample ◽  
Size Distortions ◽  
The Impact

In this paper, we present a new asymptotically normal test for out-of-sample evaluation in nested models. Our approach is a simple modification of a traditional encompassing test that is commonly known as Clark and West test (CW). The key point of our strategy is to introduce an independent random variable that prevents the traditional CW test from becoming degenerate under the null hypothesis of equal predictive ability. Using the approach developed by West (1996), we show that in our test, the impact of parameter estimation uncertainty vanishes asymptotically. Using a variety of Monte Carlo simulations in iterated multi-step-ahead forecasts, we evaluated our test and CW in terms of size and power. These simulations reveal that our approach is reasonably well-sized, even at long horizons when CW may present severe size distortions. In terms of power, results were mixed but CW has an edge over our approach. Finally, we illustrate the use of our test with an empirical application in the context of the commodity currencies literature.

scholarly journals Z-score vs minimum variance preselection methods for constructing small portfolios

2020 ◽  
Vol 17 (1) ◽  
pp. 64-76
Author(s):  
Francesco Cesarone ◽  
Fabiomassimo Mango ◽  
Gabriele Sabato
Keyword(s):  
Risk Reduction ◽  
Real World ◽  
Minimum Variance ◽  
Portfolio Performance ◽  
Z Score ◽  
Real World Data ◽  
World Data ◽  
Preliminary Results ◽  
Out Of Sample ◽  
Variance Approach

Several contributions in the literature argue that a significant in-sample risk reduction can be obtained by investing in a relatively small number of assets in an investment universe. Furthermore, selecting small portfolios seems to yield good out-of-sample performances in practice. This analysis provides further evidence that an appropriate preselection of the assets in a market can lead to an improvement in portfolio performance. For preselection, this paper investigates the effectiveness of a minimum variance approach and that of an innovative index (the new Altman Z-score) based on the creditworthiness of the companies. Different classes of portfolio models are examined on real-world data by applying both the minimum variance and the Z-score preselection methods. Preliminary results indicate that the new Altman Z-score preselection provides encouraging out-of-sample performances with respect to those obtained with the minimum variance approach.

Sign in / Sign up

Export Citation Format

Share Document