Unraveling S P500 stock volatility and networks – An encoding and decoding approach

by   Xiaodong Wang, et al.

We extend the Hierarchical Factor Segmentation(HFS) algorithm for discovering multiple volatility states process hidden within each individual S P500 stock's return time series. Then we develop an associative measure to link stocks into directed networks of various scales of associations. Such networks shed lights on which stocks would likely stimulate or even promote, if not cause, volatility on other linked stocks. Our computing endeavors starting from encoding events of large return on the original time axis to transform the original return time series into a recurrence-time process on discrete-time-axis. By adopting BIC and clustering analysis, we identify potential multiple volatility states, and then apply the extended HFS algorithm on the recurrence time series to discover its underlying volatility state process. Our decoding approach is found favorably compared with Viterbi's in experiments involving both light and heavy tail distributions. After recovering the volatility state process back to the original time-axis, we decode and represent stock dynamics of each stock. Our measurement of association is measured through overlapping concurrent volatility states upon a chosen window. Consequently, we establish data-driven associative networks for S P500 stocks to discover their global dependency relational groupings with respect to various strengths of links.


page 10

page 12


Nonparametric Test for Volatility in Clustered Multiple Time Series

Contagion arising from clustering of multiple time series like those in ...

Stock Volatility Prediction using Time Series and Deep Learning Approach

Volatility clustering is a crucial property that has a substantial impac...

Bayesian prediction of jumps in large panels of time series data

We take a new look at the problem of disentangling the volatility and ju...

Seasonal FIEGARCH Processes

Here we develop the theory of seasonal FIEGARCH processes, denoted by SF...

Discovering Multiple Phases of Dynamics by Dissecting Multivariate Time Series

We proposed a data-driven approach to dissect multivariate time series i...

ESG2Risk: A Deep Learning Framework from ESG News to Stock Volatility Prediction

Incorporating environmental, social, and governance (ESG) considerations...

Time series copula models using d-vines and v-transforms: an alternative to GARCH modelling

An approach to modelling volatile financial return series using d-vine c...

Please sign up or login with your details

Forgot password? Click here to reset