Web-based attacks, such as phishing, malware distribution, credential stealing, and phishing redirections, still rely on malicious URLs. Traditional blacklist-based and rule-based approaches are ineffective in detecting emerging, ephemeral and camouflaged URLs because they rely on historical knowledge of known threats or user-defined patterns. A hybrid fusion approach based on a CNN–RF with attention mechanism is for binary URL classification, which fuses character-level lexical feature learning and structured URL feature extraction. A URL is described in two parallel streams: a sequence of characters on a lexical level represented by a Convolutional Neural Network (CNN), and an engineered structured feature vector for Random Forest (RF) baseline classification and neural feature fusion. The CNN stream extracts discriminative character patterns from URL strings, while the structured-feature stream encodes explicit statistical, lexical, protocol- and domain-oriented features, such as URL length, special-character count, numeric-character count, character-distribution entropy, top-level-domain encoding, subdomain count, phishing-indicator terms, and protocol-safety behavior. A fusion mechanism with attention functions fuses the lexical and structured representations, allowing the final classifier to learn from both implicit (character-level) and explicit (URL features) patterns. We implemented and tested the framework on a balanced 60,000-URL set extracted from a Kaggle malicious-and-benign URL dataset using a stratified hold-out evaluation strategy. The accuracy, precision, recall, F1-score, and ROC-AUC were used to measure the performance. The fusion model outperformed the standalone CNN and RF models, achieving 99.90% accuracy, 99.80% precision, 100% recall, 99.90% F1-score, and 99.95% ROC-AUC values under the same experimental conditions. To facilitate real-time URL inference, the trained model was also integrated with an API service based on Flask to enable a service-oriented URL-prediction workflow.
Attention mechanism, Convolutional Neural Network, cybersecurity, feature engineering, Flask API, hybrid learning, malicious URL detection, phishing detection, Random Forest, real-time URL classification, URL security.
International Journal of Trend in Scientific Research and Development - IJTSRD having
online ISSN 2456-6470. IJTSRD is a leading Open Access, Peer-Reviewed International
Journal which provides rapid publication of your research articles and aims to promote
the theory and practice along with knowledge sharing between researchers, developers,
engineers, students, and practitioners working in and around the world in many areas
like Sciences, Technology, Innovation, Engineering, Agriculture, Management and
many more and it is recommended by all Universities, review articles and short communications
in all subjects. IJTSRD running an International Journal who are proving quality
publication of peer reviewed and refereed international journals from diverse fields
that emphasizes new research, development and their applications. IJTSRD provides
an online access to exchange your research work, technical notes & surveying results
among professionals throughout the world in e-journals. IJTSRD is a fastest growing
and dynamic professional organization. The aim of this organization is to provide
access not only to world class research resources, but through its professionals
aim to bring in a significant transformation in the real of open access journals
and online publishing.