Machine Learning Models for Classification of Sensitive Financial Documents

This research proposes a machine learning-based approach for the automated classification of sensitive financial documents. The proposed method utilizes Support Vector Machines (SVM), Naive Bayes, and a rule-based classifier to classify categorized financial documents accurately. The system extracts relevant features from the documents and trains the models to classify them into predefined categories. The performance of each model is evaluated, and the results are compared to identify the best-performing model. The result shows that all methods achieved satisfactory results in terms of accuracy, F-Score, and execution time. The experimental results show that the Naïve Bayes classifier outperforms the other models, achieving an accuracy of 95% in classifying the financial documents. The total time consumption, which takes into account manual handling of missorts and classification of files that could not be sorted, was significantly lower for the Naive Bayes implementations. The proposed method can significantly reduce the time and effort required for manual classification and can improve the accuracy of the classification process. The version with a tolerance level performed best and showed higher precision, but lower accuracy. This research can have significant implications in the financial industry, where the automated classification of sensitive financial documents can improve operational efficiency and reduce the risk of errors.