基於大語言模型提示技術與使用者輪廓的規格敏感性商品推薦研究

Abstract

隨著大語言模型(Large Language Models)的發展，LLM展現出強悍的語言理解以及推理能力，因此近期越來越多研究關注應用LLM於個人化的推薦系統。相關的研究分析使用者(user)和產品(item)間的互動，利用多樣的模型來建立個人化的系統。然而，隨著影響使用者行為的因數增多，如複雜的產品規格、個人偏好差異以及非結構化資料的呈現等，推薦的難度也會加大。此外，研究如何分析使用者的隱性偏好來推薦對規格敏感的產品，仍有待進一步探索。本論文提出一個新穎的框架，利用基於提示的策略分析使用者針對互動產品的評論以及產品規格資訊，分別用於建立使用者輪廓以及產品輪廓。憑藉大語言模型卓越的理解能力，這些檔案可以細微地捕捉到使用者的隱性偏好，並藉此在下游任務中進行推薦（如評分預測和產品推薦）—這些任務對提高個人化體驗至關重要。實驗結果顯示，本論文所提出的框架在評分預測任務中表現優異，能夠有效處理產品的複雜規格和使用者偏好，並且在對話式推薦中能根據對話中提供的資訊進行有效的產品推薦。

Recently, there has been an increasing focus in research on the potential applications of large language models (LLMs) for personalized recommendations. Previous studies utilize LLMs to analyze the interaction between users and products to establish various personalized recommendation systems. However, recommendation becomes particularly challenging when items are associated with varied attributes, influenced by personal preferences, and described primarily through unstructured data. Moreover, analyzing implicit user preferences with product specifications for specification-sensitive recommendations remains largely unexplored. In this paper, we propose a framework that fully leverages prompting-based strategies to analyze user reviews and item attributes for the generation of user and product profiles, respectively. These profiles capture users' implicit preferences and enable rating prediction or product recommendation, which are crucial for personalized recommendations. Experimental results show that our proposed framework effectively handles complex item attributes and user preferences to achieve promising performances in rating prediction and can retrieve candidate products based on dialogues in conversational recommendations well.