Author(s): Indraneel Parthasarathy, Tanu Priya and Moayed Daneshyari

Abstract: In a time when daily expenses frequently compound in the back-ground, sabotaging financial goals and independence, the Financial Decision Helper App is a game changer for personal finance. To tackle the common issue of unaccounted-for daily money dissipations and inconsistent account tracks, this model makes use of the state-of-the-art AI model, DeepSeek-R1-Distill-Llama-8B large language model fine-tuned with Low-Rank Adaptation (LoRA). The app offers personalized, step-by-step guidance on how users should spend their money, which includes clear instructions on whether something is "APPROVED," "DELAYED," "DENIED," "FORCEFULLY APPROVED," or "RECONSIDERED"-all with detailed explanations behind each decision to help people understand what impact it may have on their finances. User profile, budget cycle and transaction details are overlaid, producing financial awareness and sustainable habits. A well-designed manually constructed data set based on real-world financial situations guarantees that the model is accurate and relevant. The application is equipped with advanced strategies, including parameter-efficient fine-tuning, to effectively align the language model with the application requirements. Successive evaluation metrics such as the Precision, Recall, F1-score and BERTScore prove the effectiveness of the model in terms of predicting recommendations and generating comprehensible reasoning output. The application is designed to simplify financial management and to educate individuals by linking daily expenditures with long-term financial well-being.

DOI: 10.33545/27076571.2025.v6.i2b.185

Pages: 115-122 | Views: 91 | Downloads: 26

Download Full Article: Click Here