← Back to the feed
Energy & materialsResearch findings

LACRM · LLM-Guided Framework for Detecting Directional Competition Relationships

Original published 2026.08.26 18:06 KSTAdded to Blackdeck 2026.10.19 23:26 KST

What has changed?

Role-separated profiling with LLM enables detection of directional competition and continuous intensity modeling

Key points

  1. Detects directional competition using role-separated LLM profiling
  2. Models competition intensity via listwise learning-to-rank aligned with temporal auction behavior
  3. Outperforms strong directed GNN baselines in offline benchmark experiments

Connect it to your work

Effective in capturing directional and multifaceted competition dynamics

What remains unproven

This research note is based on an abstract; check the original for test conditions and application scope.

Originals used for this note

Links open the original publisher in a new tab.

  1. arxivLACRM: LLM-Guided Mining of Directed Advertiser Competition Relationships in E-Commerce Advertising