Enterprise Performance Prediction Based on FASSA Elman Neural Network

The Elman Neural Network, known for its distinctive architecture with recurrent connections, is a tool for dynamic modeling and capturing temporal dependencies. To enhance its capabilities, this paper introduces an innovative optimization approach, addressing challenges related to random weight initialization. We propose an enhanced Elman Neural Network optimization strategy using the Sparrow Search Algorithm (SSA), recognizing its global search capabilities and improved precision. To overcome SSA's drawbacks, a novel Firefly Optimized Sparrow Search Algorithm is introduced, integrating the Firefly Algorithm to enhance optimization complexity. The FA-SSA optimized Elman Neural Network is applied to predict enterprise performance, showcasing benefits in capturing temporal dynamics and intricate patterns. In summary, the proposed Firefly Optimized Sparrow Search Algorithm offers a robust solution for systematic weight and threshold optimization, enhancing the Elman Neural Network's performance.