Event(s) on December 2021

• Friday, 10th December, 2021

  Title:	How much can machines learn finance from Chinese text data?
  Speaker:	Prof Jianqing Fan, Princeton University, USA
  Time/Place:	11:00  -  12:00 Zoom, (Meeting ID: 973 9570 8109)
  Abstract:	Most studies on equity markets using text data focus on English-based specified sentiment dictionaries or topic modeling. However, can we predict the impact of news directly from the text data? How much can we learn from such a direct approach? We present here a new framework for learning text data based on the factor model and sparsity regularization, called FarmPredict, to let machines learn financial returns automatically. Unlike other dictionary-based or topic models that have stringent pre-screening processes, our framework allows the model to extract information more fully from the whole article. We demonstrate our study on the Chinese stock market, as Chinese text has no natural spaces between words and phrases and the Chinese market has a very large proportion of retail investors. These two specific features of our study differ significantly from the previous literature that focuses on English-text and the U.S. market. We validate our method using the literature on the Chinese stock market with several existing approaches. We show that positive sentiments scored by our FarmPredict approach generate on average 83 bps stock daily excess returns, while negative news has an adverse impact of 26 bps on the days of news announcements, where both effects can last for a few days. This asymmetric effect aligns well with the short-sale constraints in the Chinese equity market. As a result, we show that the machine-learned sentiments do provide sizeable predictive power with an annualized return of 116% with a simple investment strategy and the portfolios based on our model significantly outperform other models. This lends further support that our FarmPredict can learn the sentiments embedded in financial news. Our study also demonstrates the far-reaching potential of using machines to learn text data.

  
  

• Wednesday, 15th December, 2021

  Title:	Change Detection, Estimation, and Segmentation
  Speaker:	Prof David Siegmund, Stanford University, USA
  Time/Place:	09:00  -  10:00 Zoom, (Meeting ID: 966 6932 6328)
  Abstract:	Beginning with a very brief survey of three remarkable papers in the Amer. J. Math. in 1939 by authors S.O. Rice, H. Hotelling, and H. Weyl and the papers on change-point detection in the 1950s by E.S. Page, I will discuss the maximum score statistic to detect and estimate change-points in the level, slope, or other nonlinearity of an otherwise simple regression function and to segment the sequence when there appear to be multiple changes. Sequential detection of change-points, especially slope changes, will also be discussed. Examples involving temperature variations, levels of atmospheric greenhouse gases, suicide rates, incidence of Covid-19, and excess deaths during the Covid-19 pandemic illustrate the general theory. Motivated by applications that include fMRI analysis and clustering problems of astrophysics, I will also review methods for detecting "bumps" in random fields. Combining these methods leads to the study of spatio-temporal processes, where the spatial features can be either (A) unstructured vectors of observations or (B) random fields where changes of interest are expected to be geometrically clustered. Our goals can be either (C) to detect and estimate the position of local changes in the random fields arising from the temporal structure or simply (D) to determine that there are changes in the random fields, without identifying their number or location. Aspects of this research involve collaboration with Fang Xiao, Li Jian, Liu Yi, Nancy Zhang, Benjamin Yakir, Li (Charlie) Xia, and the late Keith Worsley.

All years