Factor Clustering with t-SNE

56 Pages Posted: 27 Oct 2020

See all articles by Philip Greengard

Philip Greengard

affiliation not provided to SSRN

Yukun Liu

University of Rochester - Simon Business School

Stefan Steinerberger

University of Washington

Aleh Tsyvinski

Yale University - Cowles Foundation

Date Written: September 20, 2020

Abstract

We cluster asset pricing factors using the t-distributed Stochastic Neighborhood Embedding (t-SNE), one of the most empirically successful dimensionality reduction techniques. t-SNE endogenously separates the strategies into six distinct clusters. The first five clusters resemble the standard value, momentum, investment, profitability, and volatility strategies. The sixth cluster is new, and we denote it as the Firm cluster. We show that the first five clusters are low dimensional and are dominated by their corresponding first principal components, while the Firm cluster is intrinsically high dimensional.

Suggested Citation

Greengard, Philip and Liu, Yukun and Steinerberger, Stefan and Tsyvinski, Aleh, Factor Clustering with t-SNE (September 20, 2020). Available at SSRN: https://ssrn.com/abstract=3696027 or http://dx.doi.org/10.2139/ssrn.3696027

Philip Greengard

affiliation not provided to SSRN

Yukun Liu (Contact Author)

University of Rochester - Simon Business School ( email )

Rochester, NY 14627
United States

Stefan Steinerberger

University of Washington ( email )

box 354350
Seattle, WA 98195-4350
United States

Aleh Tsyvinski

Yale University - Cowles Foundation ( email )

28 Hillhouse Ave
New Haven, CT 06520-8268
United States
203-432-9163 (Phone)

Do you have a job opening that you would like to promote on SSRN?

Paper statistics

Downloads
146
Abstract Views
731
rank
241,317
PlumX Metrics