Classification of Bank Statement using CNN Algorithm for Accounting Support

Accounting operations involve accurate and timely processing of financial documents Efficiently. The bank statements are classified manually in conventional method are time consuming causes several errors and labour-intensive process in large volume of data. The research study proposed convolutional neural network model for automated classification of bank statement and categorization based on the accounting workflow. The study utilizes text-based and image-based representation of bank statement that enable structural and spatial features. The system uses dataset for classification, input processing followed by CNN classification. The proposed model is trained with labelled dataset collected from the financial institution that covers layouts and formats. The results of the experiment provide enhanced accuracy and outperform other models that achieve robustness. The accounting system is inte4grated with CNN model reduces operational cost, improved efficiency of audit and streamlined financial entry. The research also highlights the potentiality of neural network for automated classification bank statement and documents in the accounting domain. In this study, accuracy was slightly lower for dataset 3 is $82 \%$ but significantly higher for dataset $\mathbf{2}$ is $\mathbf{9 9 \%}$.

Ҳали таржима қилинмаган