Statistical Machine Learning for Financial Prediction and Inference
Abstract
In this lecture, the speaker will introduce recent developments of statistical machine learning methods for analysis of Big Data in Finance. Motivated by stylized features such as heavy tails and cross-sectional dependence, he introduces a simple method for dependence adjustment and robustfication principles for dealing with heavy-tailed Big-Data issues in financial data. He then applies factor models to extract latent factors for prediction, Factor-Adjusted Robust Multiple test (FARM-test) and model selection (FARM-select), and to complete large covariance matrices for high-frequency financial data. He provides new statistical machine learning methods that extract latent factors that can partially be explained by several observed explanatory proxies such as the Fama-French factors in financial returns and consumption-wealth variables, financial ratios, yield spreads, and term structures in diffusion index construction. These latent factors are then further combined to predict economic and financial outcomes, such as bond risk premia and high-frequency financial returns, via a multi-index model. Empirically, he applies the model to combine high-frequency signals for predicting financial returns, to forecast US bond risk premia, and to test Fama-French factor models.
About the speaker
Prof. Fan Jian-Qing received his PhD in Statistics from the University of California, Berkeley in 1989. He has been appointed as Assistant, Associate, and Full Professor at the University of North Carolina at Chapel Hill from 1989 to 2003, Professor at the University of California, Los Angeles from 1997 to 2000, Professor of Statistics at the Chinese University of Hong Kong from 2000 to 2003. He joined Princeton University in 2003, where he is currently Frederick L. Moore '18 Professor of Finance, Professor of Statistics and of Operations Research and Financial Engineering.
Prof. Fan’s research interests are in statistical methods in finance, economics, risk management, high-dimensional statistics and machine learning computational biology, biostatistics, nonparametric modeling, longitudinal and functional data analysis, nonlinear time series, wavelets and their applications, among others. He has coauthored two highly-regarded books – Local Polynomial Modeling and Its Applications and Nonlinear Time Series: Nonparametric and Parametric Methods, and authored or coauthored over 170 articles on finance, economics, computational biology, semiparametric and non-parametric modeling, statistical learning, nonlinear time series, survival analysis, longitudinal data analysis, and other aspects of theoretical and methodological statistics. He has been consistently ranked as a top 10 highly-cited mathematical scientist since the existence of such a ranking.
Prof. Fan received numerous prestigious awards including the Presidents' Award given by the Committee of Presidents of Statistical Societies, the Humboldt Research Award for lifetime achievement, the Morningside Gold Medal of Applied Mathematics, the Pao-Lu Hsu Award by the International Chinese Statistical Association and the Guy Medal in Silver by the Royal Statistical Society. He was an invited speaker at the 2006 International Congress for Mathematicians. He is a Fellow of the American Association for the Advancement of Science, the Institute of Mathematical Statistics, and the American Statistical Association. He was elected as Academician of Academia Sinica in 2012.
