Skip to content
All library documents

Learning Firm Linkages for Momentum Spillover Strategies

Article arXiv papers · Author: Ryan Samson et al.
Read the original

Summary

This document proposes Characteristic Vector Linkages as a proxy for relationships among firms. It uses characteristics to estimate linkages, first by measuring Euclidean similarity and then by applying Quantum Cognition Machine Learning to learn similarity. The goal is to represent how firms are connected in a way that can inform momentum spillover trading strategies.

The reported results say that strategies built with both similarity methods can be profitable, and that the learned QCML similarity performs better than the Euclidean version. The description does not identify the data, sample period, transaction costs, risk controls or evaluation design, and provides no numerical performance measures. The comparison therefore supports the proposed method within the study as described, but leaves its robustness and applicability across markets unclear.

Key ideas

  • Characteristic Vector Linkages are proposed as a proxy for firm relationships.
  • Euclidean distance and QCML are used to estimate similarity between firms.
  • The learned similarity measure is applied to momentum spillover strategies.
  • The document reports stronger strategy performance with QCML similarity than with Euclidean similarity.
  • Details needed to assess costs, robustness and out-of-sample performance are not provided.

Tags

EquitiesMomentumMachine learningStatisticsFactor investing
Full text
# Supervised Similarity for Firm Linkages


# Supervised Similarity for Firm Linkages









We introduce a novel proxy for firm linkages, Characteristic Vector Linkages (CVLs). We use this concept to estimate firm linkages, first through Euclidean similarity, and then by applying Quantum Cognition Machine Learning (QCML) to similarity learning. We demonstrate that both methods can be used to construct profitable momentum spillover trading strategies, but QCML similarity outperforms the simpler Euclidean similarity.

Shown in full with attribution under the source's licence. Licence: abstract CC0

This summary was written by Stratmill's research agent from the original; it is not a copy of the source.