This repository aims to provide information (data, codes) of my current working papers in investment management, computational econometrics, data science, artificial intelligence, and text mining/webscraping.
Further studies should pursue applications to asset allocation under conditions of market distress (tail-risk hedging strategy) or to a user-specified scenario (calibrate a financial portfolio to the goals and risk tolerance of the user). In addition, I want to identify impactful news articles and construct news factor. One can apply machine learning techniques to identify relevant tags for a story and to understand which patterns can impact the market most.
- Fernando A.B. Sabino da Silva and Flavio A. Ziegelmann. Robust Portfolio Optimization with Multivariate Copulas: a Worst-Case CVaR approach.
Software: R, Matlab.
Keywords: Asset Allocation; Finance; Gaussian Copula; Linear Programming; Mixed Copula; Risk Management; S&P 500; Scenarios; WCCVaR.
JEL Codes: G11; G12; G17.
Data used: Daily data of adjusted closing prices of all stocks that belong to S&P 500 market index from July 2st, 1990 to December 31st, 2015.
- Fernando A.B. Sabino da Silva, Flavio A. Ziegelmann, Joao F. Caldeira. Mixed Copula Pairs Trading Strategy on S&P 500.
Software: Matlab, R.
Keywords: Pairs Trading; Distance; Copula; Long-Short; Quantitative Strategies; S&P 500; Statistical Arbitrage.
JEL Codes: G11; G12; G14.
Data used: Daily data of adjusted closing prices of all stocks that belong to S&P 500 market index from July 2st, 1990 to December 31st, 2015.
Press Coverage: Quantpedia.
- Fernando A.B. Sabino da Silva, Tainan B.F. Boff. Pairs Trading Strategy Combining Copula and Random Forest.
Software: R, Matlab.
Keywords: Pairs Trading; Copula, Distance, Long- Short; Quantitative Strategies; Statistical Arbitrage, High-Frequency, Realized Variance, Random Forest.
JEL: G11, G12, G14.
Data used: We select 41 assets that had non-zero returns in more than eighty percent of the time intervals at a sampling frequency of 5 minutes. All stocks belong to BMF&BOVESPA (Brazil) from January 2nd, 2010 to December 31st, 2016.
- Hudson C. Costa, Fernando A. B. Sabino da Silva, Sabino P. da Silva Junior. Macroeconomic Forecasting: Optimizing via Natural Language Processing and Text Mining Analytics.
Software: R.
Data used: Web scraping minutes of the Copom (Monetary Policy Committee of the Central Bank of Brazil - https://www.bcb.gov.br/?id=MINUTES) from 2000 to 2018. Auxiliary data is collected from the Quandl API (public data platform from around the world).
- Fernando A. B. Sabino da Silva. Additive nonparametric regression estimation via backfitting and marginal integration: Small sample performance.
Software: Originally, the codes had been written in Gauss. The attached code written in R is a modified version, including Smooth Backfitting Estimator (SBE).
Key Words: Additive nonparametric regression, Local polynomial estimation, Automatic bandwidth selection, Backfitting estimation, Marginal integration.
JEL Code: C14, C15.
*Data used: Simulated.
- Econometrics. Diagnostics.
Software: R.
*Data used: Simulated and downloaded from Professor Kenneth French's data library website (http://mba.tuck.dartmouth.edu/pages/faculty/ken.french/Data_Library/det_st_rev_factor_daily.html).
- Data Science. Data Science Homework Solutions.
Software: R.
*Data used: Simulated and official datasets for the Medicare.gov downloaded from (https://data.medicare.gov/data/physician-compare).
-
Introductory Statistics I (in portuguese). Intro Statistics I.
-
Introductory Statistics II (in portuguese). Intro Statistics II.
Software: R, Latex.
** Note **: Any code and data available here can be used as long as it is referenced. Please contact me at dasilvafbs@gmail.com or create an issue (https://github.com/FernandoSdaSilva/papers/issues?status=new&status=open) if you have any questions or suggestions or want any additional material. As soon as possible I will try to solve and leave comments so that other people also have access to the requested information.