Doubly Regularized Portfolio with Risk Minimization

Authors: Weiwei Shen, Jun Wang, Shiqian Ma

AAAI 2014 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental To assess the new portfolio, we apply standard evaluation criteria and conduct extensive experiments on well-known benchmarks and market datasets. Compared with various state-of-the-art portfolios, the proposed portfolio demonstrates a superior performance of having both higher risk-adjusted returns and dramatically decreased transaction volumes.
Researcher Affiliation Collaboration Applied Physics and Applied Mathematics, Columbia University, New York, NY, USA Business Analytics and Mathematical Sciences, IBM T. J. Watson Research, Yorktown Heights, NY, USA Systems Engineering and Engineering Management, The Chinese University of Hong Kong, Hong Kong
Pseudocode Yes Algorithm 1 Doubly Regularized Portfolio (DRP)
Open Source Code No The paper does not provide concrete access to source code for the methodology described.
Open Datasets Yes The first type of datasets is from the well-known academic benchmarks called Fama and French datasets (Fama and French 1992). The second type of datasets represents three popular financial asset classes, exchange-traded funds (ETF), major world stock market indices (INDEX), and individual equities sampled from the large-cap segment of the Russell 200 index (EQUITY). The data of the prices are all crawled from Yahoo! Finance in a weekly base from 2008 to 2012.
Dataset Splits Yes To avoid overfitting, we split the entire dataset into two parts. The first data points, denoted as {R +1, , R0}, are used as the training data to estimate the initial covariance matrix. The sequential portfolio allocation starts from t = 0 and lasts a total number of T time periods. We apply the rolling-horizon procedure (De Miguel et al. 2009) to sequentially perform portfolio allocation and evaluate the performance. Specifically, we shift the estimation window along the time axis by including the data point of the next period and dropping the data point of the earliest period. This rolling-horizon procedure is repeated until the end of the dataset. We set the length of the estimation window as = 120 (De Miguel et al. 2009), which means that the previous 120 data points are used to make the current decisions of portfolio allocation.
Hardware Specification No The paper does not provide specific hardware details used for running its experiments.
Software Dependencies No The paper mentions "commercial software toolbox TOMLAB/SNOPT developed by the Stanford Systems Optimization Laboratory (SOL)" but does not provide a specific version number for this or any other software dependencies, which is required for reproducibility.
Experiment Setup Yes We set the length of the estimation window as = 120 (De Miguel et al. 2009), which means that the previous 120 data points are used to make the current decisions of portfolio allocation. For the parameters in DRP such as λ1 and λ2, cross validation is applied to determine the optimal values.