📞 +91-7667918914 | ✉️ iarjset@gmail.com
International Advanced Research Journal in Science, Engineering and Technology
International Advanced Research Journal in Science, Engineering and Technology A Monthly Peer-Reviewed Multidisciplinary Journal
ISSN Online 2393-8021ISSN Print 2394-1588Since 2014
IARJSET aligns to the suggestive parameters by the latest University Grants Commission (UGC) for peer-reviewed journals, committed to promoting research excellence, ethical publishing practices, and a global scholarly impact.
← Back to VOLUME 13, ISSUE 11, NOVEMBER 2026

Spendwise: An Integrated Machine Learning Framework for Expense Categorization, Anomaly Detection, and Monthly Forecasting

Harshavardhan G, Deepak B T, Prawin Kumar C, Dr. Beenarani Manoj

👁 13 views📥 5 downloads
Share: 𝕏 f in ✈ ✉
Abstract: Personal expense management involves organizing transactions, identifying unusual spending behavior, and estimating future expenditure. This paper presents SpendSmart, a web-based expense analytics prototype integrating three machine-learning capabilities: transaction categorization, anomaly detection, and monthly expense forecasting. The system employs term frequency–inverse document frequency (TF-IDF) with Logistic Regression for text-based categorization, Isolation Forest for anomaly detection, and Linear Regression for forecasting. A React and TypeScript frontend communicates with a Python FastAPI inference service through an Express.js backend. The system was evaluated using synthetic datasets containing 1,800 transactions, 1,500 categorization examples, and 600 anomaly- evaluation records. On the categorization test benchmark, the word-plus-character candidate achieved 93.59% accuracy and a six-category macro-F1 score of 96.54%; however, candidate rankings varied across validation and test data, so the production model was retained. Anomaly detection experiments revealed substantial false-positive challenges. In forecasting, a previous-complete-month baseline outperformed the production Linear Regression model on the evaluated validation and final-test periods. These findings emphasize benchmark design, uncertainty handling, chronological evaluation, and cautious model selection. Because all datasets are synthetic and the final forecasting test covers only one distinct month, the results do not establish real-world financial performance.

Keywords: Expense analytics; transaction categorization; anomaly detection; expense forecasting; machine learning; Isolation Forest; rolling-origin evaluation.

How to Cite:

[1] Harshavardhan G, Deepak B T, Prawin Kumar C, Dr. Beenarani Manoj, “Spendwise: An Integrated Machine Learning Framework for Expense Categorization, Anomaly Detection, and Monthly Forecasting,” International Advanced Research Journal in Science, Engineering and Technology (IARJSET)

Creative Commons License This work is licensed under a Creative Commons Attribution 4.0 International License.