SmartShield NLP: Context-Aware Threat Severity Prediction System

Harsh Sadariya; Mohit Bhadla

doi:10.71097/IJSAT.v17.i2.11053

SmartShield NLP: Context-Aware Threat Severity Prediction System

Author(s)	Harsh Sadariya, Dr. Mohit Bhadla
Country	India
Abstract	Traditionally the binary classification models used by cybersecurity systems merely describe threats as either malicious or benign. Although useful at the early stages of filtering, such binary methods do not necessarily provide the picture of the context and seriousness of the threats, and thus it becomes difficult to prioritize incident response and resource allocation. Phishing in the modern shifting threat environment is one of the most prevalent and successful areas of cyber intrusion including both simple and innocent spam and much more serious and malicious attacks with an objective to commit either credentials theft or to deploy ransomware. To address this deficiency, this study seeks to present Beyond Binary: NLP-Based Threat Severity Prediction to Enhanced Security Response that uses Natural Language Processing (NLP) to scan the phishing emails and subsequently categorizes the emails based on their respective levels of threat rather than a simple safe/unsafe result. Phishing email datasets (enron and phish tank) are used as experimental data set in the study; the datasets are a rich source of real-life textual data. The emails are preprocessed using Python and Google Colab as a development environment through tokenization, removal of stopwords, lemmatization and embedding which are implemented through TF-IDF and transformer-based embeddings (BERT). Next, machine learning and deep learning models (Random Forest, XGBoost, LSTM and Transformer architectures) are trained to predict the levels of severity in the categories of low, medium, high and critical. This multi-class severity prediction goes beyond the binary system of phishing detection, providing high-resolution information about the possible effects of the threat. The suggested framework enhances security response mechanisms due to the fact that it enables prioritization of threats in addition to intelligent management of alerts. Threats of low severity can be automatically filtered, medium-level phishing tackles limited to review and high-to-critical level threats sent to the incident response teams to act immediately. This results in an increased usage of resources and reduces the problem of alert fatigue within Security Operations Centers (SOCs). This study is expected to produce a more scalable NLP-driven model, with the capacity to detect the existence of phishing attacks and respond to them in a context-aware and severity-based manner to offer a higher level of resilience to the current cyber attacks.
Keywords	NLP, Phishing Emails, Threat Severity Prediction, Cybersecurity, Machine Learning.
Field	Engineering
Published In	Volume 17, Issue 2, April-June 2026
Published On	2026-05-15
DOI	https://doi.org/10.71097/IJSAT.v17.i2.11053

View / Download PDF File

doi

CrossRef DOI is assigned to each research paper published in our journal.

IJSAT DOI prefix is
10.71097/IJSAT

Downloads

Research Paper Format Copyright Permission Form and Undertaking Form Cover Page Vol 17 Isu 2 Cover Page Vol 17 Isu 1 Cover Page Vol 16 Isu 4

All research papers published on this website are licensed under Creative Commons Attribution-ShareAlike 4.0 International License, and all rights belong to their respective authors/researchers.

CC-BY-SA

About IJSAT Fees & Payment Current Issue Publication Archive	Submit Research Paper Track Submission Status Publication Guidelines Publication Ethics Peer Review & Plagiarism	Join as a Reviewer Editors & Reviewers Reviewer Referral Program Get Reviewer Membership Certi.	Website/Journal Policies Usage Policy Content Policies Privacy Policy

Contact Us	Message on WhatsApp	+91-9687-182-185	editor@ijsat.org

International Journal on Science and Technology

A Widely Indexed Open Access Peer Reviewed Multidisciplinary Bi-monthly Scholarly International Journal

SmartShield NLP: Context-Aware Threat Severity Prediction System

Share this