Sustainable Accounting and Classification of Account Code with Deep Learning Models
Abstract
Organizations strive for financial management along with environmental and social goals to achieve sustainable accounting. Even though. It faces several challenges in account code mapping and accurate classification in the complex financial system, in which misclassification and inconsistency hinder sustainable accounting and transparency. The study proposed a deep learning (DL) framework for the classification of account codes for sustainable accounting. The algorithms include SqueezeNet, Inception V3, DeepLoc, VGG-16, Painters, and VGG-19. The study integrates transfer learning and feature extraction techniques for the reduction of manual dependency and enhanced classification accuracy. The findings of the experiments show that high accuracy is achieved with the InceptionV3 Net. Real-time application provided efficiency with Squeeze-Net. The experimental results reveal that deep learning models improve reliability in classification, which leads to sustainable accounting. It reduces human error and improves resource optimization. The study emphasizes DL advancement in the accounting domain. The study achieved an accuracy of 98% in the classification of accounting codes.