Large Language Model Data Query Method Based on Feedback Optimization Prompt Word Engineering

With the rapid development of large language models (LLMs) and their remarkable capabilities in conversational understanding and text generation, significant achievements have been made in the field of general question answering. However, in specific domains, these models often fail to fully leverage their strengths. To address this, this paper propose a Q&A-style database query and statistical analysis method that integrates AI-powered LLM technology, aiming to optimize data retrieval processes for corporate production operations and project management. The core of this method lies in first performing a structured parsing of user-input natural language questions to accurately capture query intent and extract initial data requirements. Subsequently, by leveraging advanced prompt engineering techniques, the large language model's understanding of specific database schemas and field tables is significantly enhanced. This technology enables the intelligent conversion of natural language query conditions into efficient database query statements, which are automatically executed to precisely retrieve target data. Furthermore, a feedback optimization mechanism based on LLM prompt engineering is introduced, which iteratively refines prompt strategies based on user feedback. This ensures that the model can generate query expressions that more accurately align with user expectations, thereby substantially improving the accuracy and practicality of queries. This approach not only significantly enhances the model's ability to comprehend field tables within specific database environments but also provides robust support for the intelligent transformation of production operations and project management.