What question did this study set out to answer?

The aim is to develop a predictive model for enterprise financial monitoring using advanced machine learning techniques.

May 22, 2026Open Access

Development and application of a system-integrated intelligent income and expenditure monitoring and analysis tool based on LSTM optimisation and dynamic sandbox projections

Key Points

The aim is to develop a predictive model for enterprise financial monitoring using advanced machine learning techniques.
Developed a multi-dimensional prediction model integrating macroeconomic indicators.
Utilized a Bayesian LSTM network for enhanced prediction accuracy.
Created a dynamic causal inference engine for real-time financial scenario simulation.
Achieved a ±7% improvement in prediction accuracy, with a 53% reduction in error compared to traditional models.
Reduced rolling loan costs for manufacturing enterprises by 23%.
Lowered inventory backlog risk in the retail industry by 3.2 million RMB.

Abstract

In this study, a multi-dimensional prediction model and visual decision support system integrating macroeconomic indicators are constructed for the dynamic monitoring needs of enterprise financial management. It achieves ±7% prediction accuracy improvement (53% error reduction compared with traditional model) by introducing Bayesian LSTM network, and develops interactive causal inference engine to support dynamic scene simulation. Empirical evidence shows that the system can reduce the rolling loan cost of manufacturing enterprises by 23% and the risk of inventory backlog in retail industry by 3.2 million RMB. The research innovations are: 1. proposing a multimodal data-driven financial health assessment system 2. establishing a flexible budget derivation algorithm that integrates policy text analysis.

اسأل الذكاء الاصطناعي

Bookmark

View Full Paper