针对推荐系统的合成数据生成方法及介质
By combining a two-stage framework of large-scale language model LLM and diffusion model, a high-quality, diverse and controllable recommendation dataset is generated, which solves the problems of fidelity, diversity and controllability of synthetic datasets in existing technologies and improves the effectiveness of recommendation systems.
CN122413348APending Publication Date: 2026-07-17NINGBO UNIV
View PDF 0 Cites -1 Cited by
Patent Information
- Application Number
- CN202610873969.2
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2026-06-17
- Publication Date
- 2026-07-17
Smart Images

Figure CN122413348A_ABST
Abstract
本发明涉及一种针对推荐系统的合成数据生成方法,包括步骤S1、初始化大语言模型LLM、扩散模型构成的合成数据生成框架,其中的扩散模型利用真实数据配合大语言模型LLM训练获取;S2、用户输入推荐指令;S3、大数据模型LLM对推荐指令进行解析,进而根据推荐指令生成数据样本作为种子样本;S4、扩散模型对种子样本进行条件扩展计算,进而获取合成数据样本。该针对推荐系统的合成数据生成方法,能生成具有代表性的种子样本,并能进行条件扩展,进而生成高质量、多样化且可控的推荐数据集。本发明还涉及一种计算机可读存储介质,其上存储有可被处理器执行而实现前述合成数据生成方法的计算机程序。
Need to check novelty before this filing date? Find Prior Art