An Intelligent Budget Management Mobile Application Based on a Recurrent Neural Network

Budgeting is important for both people and businesses to manage their finances, save, stay out of debt, and be ready for financial emergencies. This paper aims to propose an intelligent mobile application to reduce financial risks and support better ways of saving and investing through intelligent budgeting. A lot of budgeting applications were lack Arabic language support and AI-based forecasting capabilities. This gap is addressed by development of application that supports English and Arabic speakers through Recurrent Neural Network (RNN) for enhanced financial and management capabilities. The model was trained by a real-world financial dataset, preprocessing it, and transforming its data. The results demonstrated that the model was highly accurate in the classification of financial transactions based on the level of risk. The application was tested on new data; it reached an overall accuracy of 97.45%. Precision, recall, and F1 measures all higher than 0.97 for each of the risk categories: low, medium, and high. These results validate the reliability of the system, particularly in detecting critical low- and high-risk behaviors, and its capacity to help users make safer and more informed decisions regarding their finances.