Research on Agricultural Enterprise Risk Management Decision Making Based on Recurrent Neural Network (RNN)

Agricultural enterprises play a crucial role in economic development, but they inevitably encounter operational risks during the development process. Therefore, an effective business risk management system is a key link for agricultural enterprises to achieve long-term stable development. An accurate business risk prediction framework can help agricultural enterprises better cope with internal and external challenges, ensuring their sustainable development in complex and changing environments. In response to this issue, this study improved the recurrent neural network model and applied it to the operational risk management of agricultural enterprises. Specifically, we designed an improved Light Spectrum Optimizer (LSO) and optimized the two main hyperparameters of the RNN using the improved LSO algorithm. These two hyperparameters are the number of neurons in the hidden layer of the RNN framework and the learning rate. Finally, based on a test set consisting of 25 agricultural enterprises, the improved LSO based RNN (ILSO-RNN) framework was evaluated, and the results showed that the ILSO-RNN framework has high accuracy in conducting agricultural enterprise risk assessment.