Answers are grounded in this paper and cite their sources.
Simulation study demonstrates improved cost prediction accuracy and reduced deviation in manufacturing enterprises, highlighting the benefits of machine learning with dynamic feedback control.
Enterprise cost management in intelligent accounting environments requires integrated data processing, dynamic accounting, predictive control, and decision support. Traditional cost-management models suffer from fragmented data, rigid allocation rules, and weak real-time control capability. This study proposes a systematic optimization path for enterprise cost management based on multi-source data standardization, activity-based costing, regression-based cost-driver identification, machine-learning prediction, and dynamic feedback control. First, heterogeneous cost data are integrated, cleaned, and standardized to form a unified multidimensional cost dataset. Second, an intelligent activity-based costing model dynamically allocates resource costs to activities and cost objects according to observed driver variables. Third, machine-learning regression models predict future costs and identify deviations through standardized residuals, enabling rolling adjustment and early warning. Finally, multidimensional cost analysis and scenario-based optimization are used to support resource allocation decisions. Experiments using monthly cost data from manufacturing enterprises show that the proposed method reduces MSE to 58.73, MAE to 5.42, and the cost deviation rate to 0.051, outperforming traditional accounting, non-intelligent ABC, and machine-learning models without feedback adjustment. The method provides a data-driven framework for intelligent cost prediction, time-series control, and enterprise decision support.
No takes yet. Share an insight, caveat, or question.
Dong et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: