A comprehensive review of interpretable machine learning techniques for phishing attack detection - Future directions and integration approaches

Aug 3, 2026

Phishing attacks remain a persistent and fast-evolving cybersecurity threat, requiring detection methods that go beyond conventional accuracy by adding transparency and user trust. This paper reviews interpretable machine learning models for phishing detection, including rule-based approaches, decision trees, and additive explanation methods such as SHAP. Evaluation focuses on computational efficiency, prediction accuracy, and interpretability, and the work discusses integration into existing systems. It also outlines future research to improve scalability, robustness, and adaptation to emerging phishing techniques.

LanguageEnglish

CategoryResearch & Report

FormatPDF

Pages11

Reading Time28 min

AccessFree to download

Abstract

Introduction

Phishing attacks and detection challenges

Background and Related Work

Explanation of phishing attacks, types, and characteristics

Interpretable machine learning approaches for phishing detection

Rule-based models and decision trees

Additive explanation methods (e.g., SHAP)

Integration into detection systems

Enhancing functionality and user experience

Future research directions

Scalability, accuracy, adaptability, and real-time threat intelligence

A comprehensive review of interpretable machine learning techniques for phishing attack detection Pankaj Chandre, Pallavi Bhujbal, Ashvini Jadhav, Bhagyashree Dinesh Shendkar, Aditi Wangikar, Rajneeshkaur Sachdeo Department of Computer Science and Engineering, MIT School of Computing, MIT Art Design and Technology University, Pune, India | Article history:
Received Apr 25, 2024 Revised Jun 13, 2025 Accepted Jul 10, 2025 | Phishing attacks remain a significant and evolving threat in the digital landscape, demanding continual advancements in detection methodologies. This paper emphasizes the importance of interpretable machine learning models to enhance transparency and trustworthiness in phishing detection systems. It begins with an overview of phishing attacks, their increasing sophistication, and the challenges faced by conventional detection techniques. A range of interpretable machine learning approaches, including rule-based models, decision trees, and additive models like Shapley additive explanations (SHAP), are surveyed. Their applicability in phishing detection is analyzed based on computational efficiency, prediction accuracy, and interpretability. The study also explores ways to integrate these methods into existing detection systems to enhance functionality and user experience. By providing insights into the decision-making processes of detection models, interpretable machine learning facilitates human supervision and intervention, strengthening overall system reliability. The paper concludes by outlining future research directions, such as improving the scalability, accuracy, and adaptability of interpretable models to detect emerging phishing techniques. Integrating these models with real-time threat intelligence and deep learning approaches could boost accuracy while preserving transparency. Additionally, user-centric explanations and humanin-the-loop systems may further enhance trust, usability, and resilience in phishing detection frameworks.
This is an open access article under the CC BY-SA license.
| | --- | --- | | Keywords:
Cybersecurity
Decision-making processes Detection methodologies Interpretable machine learning Phishing attacks | | Corresponding Author: Pankaj Chandre Department of Computer Science and Engineering, MIT School of Computing MIT Art Design and Technology University Loni Kalbhor, Pune, India [Email: pankajchandre30@gmail.com](Email: pankajchandre30@gmail.com) Article Info ABSTRACT 1. INTRODUCTION Phishing attacks pose a significant threat to cybersecurity, targeting individuals, organizations, and critical infrastructures worldwide. These attacks use deceptive techniques to fool users into disclosing private information, including bank account information and login credentials [1] . Although traditional machine learning techniques have been used to detect phishing attempts, their interpretability and transparency issues frequently restrict their efficacy [2] . The increasing complexity of phishing techniques is driving the demand for sophisticated detection mechanisms that can identify subtle patterns of attack [3] . As a result, approaches for interpretable machine learning have surfaced as viable remedies, providing transparent models that provide light on the decision-making process [4], [5] . This study provides a thorough assessment and analysis of current methods in order to investigate the function of interpretable machine learning in phishing attack detection [6] . The introduction lays forth the goals and framework of this research, which prepares the reader for a thorough analysis of interpretable machine learning techniques for thwarting phishing attacks. Phishing attacks continue to be a major cybersecurity concern, with millions of incidents reported globally each year. According to industry reports, phishing attacks accounted for over 36% of data breaches in recent years, causing billions of dollars in financial losses for individua

FAQs

Why are interpretable machine learning models important for phishing detection?

Interpretable models make the decision process transparent, improving trust and enabling human supervision. This helps analysts understand and act on predictions during phishing detection.

Which interpretable techniques are surveyed in the paper?

The paper surveys rule-based models, decision trees, and additive explanation approaches such as SHAP. It analyzes how each supports interpretability alongside detection performance.

How does the paper evaluate these interpretable methods for phishing detection?

Evaluation considers computational efficiency, prediction accuracy, and interpretability. The goal is to balance detection effectiveness with understandable decision-making.

Aug 3, 2026

Upload by Evangeline