Large Language Model-Aided Production-Level Intelligent Economic Forecasting Method

In a rapidly changing industrial environment, the accuracy of economic forecasting is increasingly critical. With the rise of the Large language Model (LLM), a new path for economic forecasting has been opened up. Therefore, the purpose of this study is to use LLM to improve the accuracy and flexibility of industrial economic forecasting. By analyzing the shortcomings of existing forecasting methods, the Bidirectional Encoder Representations from Transformers (BERT) model is used to process economic data, and the text is transformed into high-dimensional vectors to accurately capture economic characteristics. Subsequently, it is combined with the Generative Pre-trained Transformer (GPT) large language model to deepen the data analysis, and assists the Bidirectional Long Short-Term Memory (BiLSTM) model for industrial economic prediction. In addition, the optimized Dung Beetle Optimization (DBO) algorithm is introduced to fine-tune BiLSTM parameters, and a new framework of economic intelligent prediction assisted by the large language model is constructed. The results show that this method significantly improves the accuracy, timeliness and comprehensiveness of the forecast. And it can be flexibly applied to multi-industry and multi-time scale forecasting.