An open access journal
Enhancing Telecom Fraud Prediction Accuracy Using a Combined CNN-LSTM Model with Bahdanau Attention Mechanism
Abstract
Telecom fraud Financial fraud is committed by telephone, Internet and SMS, and it is important to identify and prevent these activities. In this paper, a telecom fraud prediction model is established based on the collected telecom fraud data set. Firstly, the environment in which fraud occurs is analyzed through data visualization technology, and then the influence of bank card usage on fraud probability is discussed by using logistic regression model, and it is found that there is a significant correlation between improper use of bank card and fraud risk. Then, the relationship between transaction pattern and fraud is revealed through the correlation analysis of variables, and the complex relationship between different transaction types and fraud probability is deeply discussed. Finally, convolutional neural network and short-duration memory network model combined with Bahdanau attention mechanism were used to improve the prediction accuracy, and the accuracy of the model was up to 99%. This study not only improves the prediction ability of the model, but also shows the extensive application potential of the model, which provides important technical support and theoretical basis for the identification and prevention of telecom fraud.
Share and Cite
Article Metrics
References
- Chen Xiaolei, Xu Hui Construction of Evidence System for Telecommunications Fraud Cases [J] Network Security Technology and Applications, 2024 (5): 139-141.
- Qin Yutian The logical development and criminal control of telecom network fraud with fraud as the core charge [J/OL] Journal of Taiyuan University (Social Sciences Edition), 2024, 25 (4): 95-104.
- Zhang Hao Analysis of Ecological Governance of the "Black and Grey Industry Chain" in Telecommunications Network Fraud Crimes [J] Network Security Technology and Applications, 2024 (4): 148-150.
- Xiao Song, Huang Jiewu First order approximate knife cut correction ridge estimation in binary logistic regression models [J] Journal of Hunan University of Arts and Sciences (Natural Science Edition), 2024, 36 (2): 6-13.
- Zhang Jian, Liu Lin, Li Qian Research on Optimization of M-Score Model Based on Nonlinear Logistic Regression [J] Business Accounting, 2024 (10): 83-86.
- Liao Wen Research and application of Internet financial user churn prediction model based on data mining [D/OL] South China University of Technology, 2021.
- Zhao Zemin Research on Smart Habitat Mode and Behavior Prediction Based on Data Mining Technology [D/OL] Harbin Institute of Technology, 2021.
- Li Hao, Zhao Qing, Cui Chenzhou, etc Stellar Spectral Classification Algorithm Based on CNN and LSTM Composite Deep Model [J] Spectroscopy and Spectral Analysis, 2024, 44 (6): 1668-1675.
- Zhao Zemin Research on Smart Habitat Mode and Behavior Prediction Based on Data Mining Technology [D]. Harbin Institute of Technology, 2021.
- Li Huifeng, Li Tiecheng, Li Weixun. Application of Fuzzy Theory in the Evaluation of Communication Networks in Intelligent Substations [J]. Mechanical Design and Manufacturing, 2024 (04): 28-32+37.
- Cheng Junhan, Wang Shuli, Cai Zhiyuan. Remaining service life prediction of lithium batteries based on AE-LSTM [J]. Electrical and Energy Efficiency Management Technology, 2023 (09): 69-75.
- Liu Shengjiu, Liang Shupeng, Liu Ying, etc Property Analysis and Application Research of Hypergraph Entropy [C]//Chinese Society of Automation. Proceedings of the 2023 China Automation Conference [Publisher unknown], 2023: 6.
- Zhang Haiyang, Chen Yuming, Zeng Nianfeng, et al. Credit Card Fraud Detection Based on XGBoost and LR Fusion Model [J]. Journal of Chongqing University of Technology (Natural Sciences), 2024, 38 (03): 195-200.