Implementation of 'Smishing Detector': An Efficient Model for Smishing Detection Using Neural Network

Area of Science:

• Cybersecurity
• Machine Learning
• Artificial Intelligence

Background:

• Smishing, or phishing via text messages, poses a significant cybersecurity threat to smartphone users.
• Effective detection of smishing is crucial for protecting individuals and systems.
• Existing methods require efficient algorithms to combat evolving smishing tactics.

Purpose of the Study:

• To present and implement a Neural Network-based algorithm for smishing detection.
• To evaluate the efficiency of the Neural Network model compared to other machine learning algorithms.
• To identify and analyze the most effective features for smishing detection.

Main Methods:

• Implementation of a Neural Network model based on the 'Smishing Detector' framework.
• Utilizing Neural Networks to extract and analyze key features of smishing messages.
• Comparative analysis of the Neural Network's performance against other machine learning algorithms.

Main Results:

• The Neural Network model achieved a final accuracy of 97.40% in detecting smishing.
• Neural Networks demonstrated superior performance, outperforming other algorithms by 1.11%.
• Analysis identified specific features crucial for smishing detection, with Uniform Resource Locator (URL) being the most effective (94% accuracy).

Conclusions:

• Neural Networks are highly effective for detecting smishing attacks.
• The identified features, particularly URLs, significantly contribute to accurate smishing detection.
• This research provides an efficient algorithmic solution to a pressing cybersecurity challenge.