2018/02/06 by Javier Franco-Pedroso, Joaquín González-Rodríguez, Franco-Pedroso, Javier +8
Computer Science · Decision Sciences · Economics, Econometrics and Finance · #Complex Systems and Time Series Analysis #Computational Engineering #Computational Finance (q-fin.CP) #FOS: Computer and information sciences #FOS: Economics and business #FOS: Electrical engineering #Finance #Signal Processing (eess.SP) #Stock Market Forecasting Methods #Time Series Analysis and Forecasting #and Science (cs.CE) #electronic engineering #information engineering
paper · pdf · doi:10.48550/arxiv.1802.01861
openalex publication_date 2018/02/06 · openalex created_date 2025/10/10 · openalex updated_date 2026/07/28
In this paper, we present a novel approach to the generation of virtual\nscenarios of multivariate financial data of arbitrary length and composition of\nassets. With this approach, decades of realistic time-synchronized data can be\nsimulated for a large number of assets, producing diverse scenarios to test and\nimprove quantitative investment strategies. Our approach is based on the\nanalysis and synthesis of the time-dependent individual and joint\ncharacteristics of real financial time series, using stochastic sequences of\nmarket trends to draw multivariate returns from time-dependent probability\nfunctions preserving both distributional properties of asset returns and\ntime-dependent correlation among time series. Moreover, new time-synchronized\nassets can be arbitrarily generated through a PCA-based procedure to obtain any\nnumber of assets in the final virtual scenario. For the validation of such\nsimulated data, they are tested with an extensive set of measurements showing a\nsignificant degree of agreement with the reference performance of real\nfinancial series, better than that obtained with other classical and\nstate-of-the-art approaches.\n