quantlab.factor.predefined.literature_alpha#
Literature-backed equity alpha factors implemented as one KunQuant graph.
The bundle covers eight return-predictive characteristics from distinct economic mechanisms: the 52-week price anchor, short-term reversal, lottery preference, idiosyncratic volatility, illiquidity, gross profitability, asset growth and post-earnings-announcement drift. It is deliberately universe agnostic. A caller that researches the Nasdaq-100, S&P 500 or another point- in-time universe applies that universe to the input panel before constructing this factor.
Every characteristic has a *_raw output and a cross-sectional *_rank
output. KunQuant’s Rank ignores NaN, so a panel masked by its membership
(IndexConstituentDataset.mask) is ranked only over the valid members at each timestamp.
Inputs and windows are configurable through FactorConfig.kwargs; selected
factor_names prune unused inputs and graph branches.
References#
George and Hwang (2004), Jegadeesh (1990), Bali, Cakici and Whitelaw (2011), Ang, Hodrick, Xing and Zhang (2006), Amihud (2002), Novy-Marx (2013), Cooper, Gulen and Schill (2008), and Livnat and Mendenhall (2006).
Classes
Compute eight literature-backed equity characteristics in KunQuant. |
|
Formula settings and panel-column aliases for |