PulseAugur
EN
LIVE 06:58:55

Deep neural networks can be decomposed into linear factor representations for financial modeling

By PulseAugur Editorial · [1 sources] ·

A new research paper introduces the concept of Large and Deep Factor Models, demonstrating how deep neural networks (DNNs) trained to construct stochastic discount factors (SDFs) can be decomposed. This decomposition reveals a linear factor representation governed by the Portfolio Tangent Kernel (PTK), which effectively summarizes the network's learned features. The study shows that this PTK representation offers significant performance improvements in empirical tests using U.S. equity data, while also highlighting how increasing spectral complexity can limit finite-sample pricing capabilities. AI

IMPACT Introduces a novel method for financial modeling using deep neural networks, potentially improving quantitative finance strategies.

RANK_REASON The cluster contains a single academic paper detailing a new modeling approach. [lever_c_demoted from research: ic=1 ai=0.7]

Read on arXiv cs.LG →

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →

Deep neural networks can be decomposed into linear factor representations for financial modeling

COVERAGE [1]

  1. arXiv cs.LG TIER_1 English(EN) · Bryan Kelly, Boris Kuznetsov, Semyon Malamud, Yuan Zhang ·

    Large and Deep Factor Models

    arXiv:2402.06635v3 Announce Type: replace-cross Abstract: We show that a deep neural network (DNN) trained to construct a stochastic discount factor (SDF) admits an additive decomposition separating nonlinear characteristic discovery from the pricing rule that aggregates them. Th…

RELATED TOPICS