Behavioural Analytics for Early Detection of SIM Swap and Account Takeover in Financial Systems: A Model

The increasing adoption of digital financial services has heightened the vulnerability of financial systems to cybersecurity threats, particularly SIM swap fraud and account takeover (ATO). Conventional fraud detection mechanisms, which predominantly rely on predefined rules and transaction-based monitoring, often struggle to identify sophisticated attacks that exploit legitimate authentication channels and imitate normal customer behaviour. This study develops a Hybrid Predictive Intelligence Model (HyPIM), a behavioural analytics framework for the early detection of SIM swap fraud and account takeover in financial systems. The proposed framework integrates telecommunications-related SIM lifecycle information, device intelligence, authentication activities, geographical access patterns, and financial transaction behaviour through a five-layer architecture comprising data acquisition, behavioural feature engineering, intelligent analytics, risk assessment, and fraud detection and response. The methodology employs Random Forest, Support Vector Machine (SVM), XGBoost, and Isolation Forest algorithms to identify known fraudulent activities and emerging behavioural anomalies. The experimental evaluation utilises the SIM Swap Fraud Detection dataset and PaySim1 financial transaction dataset as complementary sources for behavioural and transactional fraud analysis. The illustrative results indicate that HyPIM achieves 99.63% accuracy, 94.80% precision, 93.10% recall, and a 0.24% false-positive rate, suggesting improved detection performance compared with individual machine learning models and conventional rule-based approaches. The study contributes an integrated approach to behavioural fraud detection by correlating SIM-related security indicators with financial account activities to support early threat identification and timely security interventions.

Authors

Institutions

Publication Details

Journal
Tech-Sphere Journal for Pure and Applied Sciences
Published
2026-11-19
DOI
https://doi.org/10.5281/zenodo.24377203
Primary Topic
Imbalanced Data Classification Techniques
Type
article
Field-Weighted Citation Impact
0.00
Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
NOV
article

Behavioural Analytics for Early Detection of SIM Swap and Account Takeover in Financial Systems: A Model

Obiageli Attoh
Tech-Sphere Journal for Pure and Applied Sciences
Imbalanced Data Classification Techniques
article

Behavioural Analytics for Early Detection of SIM Swap and Account Takeover in Financial Systems: A Model

Obiageli Attoh
article en

Abstract

The increasing adoption of digital financial services has heightened the vulnerability of financial systems to cybersecurity threats, particularly SIM swap fraud and account takeover (ATO). Conventional fraud detection mechanisms, which predominantly rely on predefined rules and transaction-based monitoring, often struggle to identify sophisticated attacks that exploit legitimate authentication channels and imitate normal customer behaviour. This study develops a Hybrid Predictive Intelligence Model (HyPIM), a behavioural analytics framework for the early detection of SIM swap fraud and account takeover in financial systems. The proposed framework integrates telecommunications-related SIM lifecycle information, device intelligence, authentication activities, geographical access patterns, and financial transaction behaviour through a five-layer architecture comprising data acquisition, behavioural feature engineering, intelligent analytics, risk assessment, and fraud detection and response. The methodology employs Random Forest, Support Vector Machine (SVM), XGBoost, and Isolation Forest algorithms to identify known fraudulent activities and emerging behavioural anomalies. The experimental evaluation utilises the SIM Swap Fraud Detection dataset and PaySim1 financial transaction dataset as complementary sources for behavioural and transactional fraud analysis. The illustrative results indicate that HyPIM achieves 99.63% accuracy, 94.80% precision, 93.10% recall, and a 0.24% false-positive rate, suggesting improved detection performance compared with individual machine learning models and conventional rule-based approaches. The study contributes an integrated approach to behavioural fraud detection by correlating SIM-related security indicators with financial account activities to support early threat identification and timely security interventions.

Tech-Sphere Journal for Pure and Applied Sciences
Dennis Osadebay University (NG)
Imbalanced Data Classification Techniques
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.

Laika access coming soon