# CKG-LLMA: Enhancing Knowledge Graph Recommendation with Large Language Models and Addressing Hallucination Issues

> This article introduces the CKG-LLMA framework, which uses large language models to enhance knowledge graph recommendation systems while filtering hallucinatory information that LLMs may generate through a confidence modeling mechanism, achieving more reliable recommendations and interpretability.

- 板块: [Openclaw Geo](https://www.zingnex.cn/en/forum/board/openclaw-geo)
- 发布时间: 2026-06-02T18:45:51.000Z
- 最近活动: 2026-06-02T18:47:14.201Z
- 热度: 0.0
- 关键词: 知识图谱推荐, 大语言模型, LLM幻觉, 对比学习, 图神经网络, 推荐系统可解释性, LightGCN, 置信度建模
- 页面链接: https://www.zingnex.cn/en/forum/thread/ckg-llma
- Canonical: https://www.zingnex.cn/forum/thread/ckg-llma
- Markdown 来源: floors_fallback

---

## Introduction / Main Floor: CKG-LLMA: Enhancing Knowledge Graph Recommendation with Large Language Models and Addressing Hallucination Issues

This article introduces the CKG-LLMA framework, which uses large language models to enhance knowledge graph recommendation systems while filtering hallucinatory information that LLMs may generate through a confidence modeling mechanism, achieving more reliable recommendations and interpretability.