Does non-linear factorisation of financial returns help build better and stabler portfolio?

  • 5 Rating
  • 1 Reviews
  • 0 Students Enrolled

Does non-linear factorisation of financial returns help build better and stabler portfolio?

Presentation for the paper "Does non-linear factorisation of financial returns help build better and stabler portfolio?"

  • 5 Rating
  • 1 Reviews
  • 0 Students Enrolled
Tags:



Courselet Content

1 courselets • 2 courselet components • 00h 45m total length

Requirements

  • SFM: Portfolio theory, Neural networks, ARMA, GARCH

Description

Abstract:

A portfolio allocation method based on linear and non-linear latent constrained conditional factors is presented. The factor loadings are constrained to always be positive in order to obtain long-only portfolios, which is not guaranteed by classical factor analysis or PCA. In addition, the factors are to be uncorrelated among clusters in order to build long-only portfolios. Our approach is based on modern machine learning tools: convex Non-negative Matrix Factorization (NMF) and autoencoder neural networks, designed in a specific manner to enforce the learning of useful hidden data structure such as correlation between the assets' returns. Our technique finds lowly correlated linear and non-linear conditional latent factors which are used to build outperforming global portfolios consisting of cryptocurrencies and traditional assets, similar to hierarchical clustering method. We study the dynamics of the derived non-linear factors in order to forecast tail losses of the portfolios and thus build more stable ones.

Paper available at:

- https://arxiv.org/abs/2204.02757

- https://papers.ssrn.com/sol3/papers.cfm?abstract_id=4076843

Quantlet: https://github.com/QuantLet/EmbeddingPortfolio

Recently Added Courses/Courselets

blog
Last Updated 3rd June 2022
  • 0
  • Free

About the Instructor

instructor
About the Instructor

Bruno Spilak is currently a research associate and PhD student at Humboldt University of Berlin with International Research Training Group IRTG1792 “High dimensional non stationary time series analysis”.

Along with a double masters degree in Statistics from Humboldt-Universität and ENSAI (a french Engineer Grande École), he has a rich background in research, scientific presentations and teaching.

 

Since 2018, he has co-taught graduate level courses related to smart data analysis, machine learning and mathematical statistics at his current institute and completed various research projects related to statistics, econometrics, machine learning and finance.

 

In parallel, Bruno has worked in multiple startups in Berlin as a Machine Learning scientist and co-built automated KYC products with deep computer vision tools.

 

Presently, he is working on the methodology to construct portfolios with non-parametric statistics tools such as neural networks.

 

Student Feedback

5
Course Rating
100%  
100%  
100%  

JB
27-06-2022
Justin Bbox

This guy likes cool. I like the way he illustrates the state-of-art staffs.