Intelligent accounting optimization method based on meta-heuristic algorithm and CNN

Area of Science:

• Accounting Technology
• Artificial Intelligence in Finance
• Computational Intelligence

Background:

• The increasing complexity of enterprise accounting necessitates advanced solutions.
• Social intelligence evolution drives the adoption of intelligent accounting practices.
• Current accounting systems require enhanced efficiency and improved accountant capabilities.

Purpose of the Study:

• To propose an intelligent accounting optimization approach.
• To enhance enterprise accounting operations and accountant capabilities.
• To provide technological support for refining accounting audit mechanisms.

Main Methods:

• Enhanced Convolutional Neural Network (CNN) framework integrating document and voucher information for multi-modal feature extraction.
• A novel method for assessing accounting quality using multi-modal accounting features to objectively evaluate financial performance.
• An optimization technique combining genetic algorithms with annealing models for accounting system improvement.

Main Results:

• The proposed approach achieved a high accuracy of 0.943.
• The mean average precision (mAP) score reached 0.812, indicating robust performance.
• Experimental results validate the effectiveness of the integrated meta-heuristic and CNN approach.

Conclusions:

• The developed intelligent accounting optimization method significantly enhances accounting audits.
• The approach offers objective financial performance evaluation and system improvement.
• This research provides a strong technological foundation for the future of intelligent accounting.