POSTER: A Multi-Signal Model for Detecting Evasive Smishing

TL;DR

This paper introduces a multi-signal detection model for identifying evasive smishing messages, combining semantic, structural, stylistic, and contextual cues, achieving high accuracy and outperforming single-signal models.

Contribution

The paper presents a novel multi-signal architecture for smishing detection that integrates diverse linguistic and structural features, improving robustness and regional awareness.

Findings

Achieved 97.89% accuracy in smishing detection

F1 score of 0.963 indicating high precision and recall

Outperformed single-stream models significantly

Abstract

Smishing, or SMS-based phishing, poses an increasing threat to mobile users by mimicking legitimate communications through culturally adapted, concise, and deceptive messages, which can result in the loss of sensitive data or financial resources. In such, we present a multi-channel smishing detection model that combines country-specific semantic tagging, structural pattern tagging, character-level stylistic cues, and contextual phrase embeddings. We curated and relabeled over 84,000 messages across five datasets, including 24,086 smishing samples. Our unified architecture achieves 97.89% accuracy, an F1 score of 0.963, and an AUC of 99.73%, outperforming single-stream models by capturing diverse linguistic and structural cues. This work demonstrates the effectiveness of multi-signal learning in robust and region-aware phishing.