Phishing Website Detection Research Articles

A new phishing-website detection framework using ensemble classification and clustering

Phishing websites are characterized by distinguished visual, address, domain, and embedded features, which identify and defend such threats. Yet, phishing website detection is challenged by overlapping these features with legitimate websites’ features. As the inter-class variance between legitimate and phishing websites becomes low, commonly utilized machine learning algorithms suffer from low performance in overlapping feature cases. Alternatively, ensemble learning that combines multiple predictions intending to address low inter-class variations in the classified data improves the performance in such cases. Ensemble learning utilizes multiple classifiers of similar or different types with multiple deviations of the training data. This paper develops a framework based on random forest ensemble techniques. The limitations of the random forest are the inability to capture the high correlation between features and their join dependency on the label. The random forest is combined with k-means clustering to capture the feature correlation. The framework is evaluated for phishing detection with a dataset of 5000 samples. The results showed the proposed framework over-performed the random forest classifier, all other ensemble classifiers, and the conventional classification algorithms. The proposed framework achieved an accuracy of 98.64%, precision of 0.986, recall of 0.987, and F-measure of 0.986.

Detecting Phishing Websites using recent Techniques: A Systematic Literature Review

The goal of this study Phishing attacks are constantly evolving, and to avoid being detected by conventional means, attackers use cutting-edge approaches. Novelty detection aims to identify previously unseen phishing attacks, including zero-day threats and sophisticated evasion tactics. Phishing attacks continue to pose significant threats to cybersecurity, exploiting human vulnerabilities and developing quickly to avoid being detected by conventional methods. In response to these challenges, this literature survey presents a comprehensive review of phishing website detection techniques, focusing on novel approaches and the latest advancements in the field. It explores dynamic analysis, real-time monitoring, and anomaly detection techniques to keep pace with the ever-changing phishing landscape. The survey addresses the persistent issue of imbalanced datasets by presenting effective strategies for handling data from significantly more legitimate websites than phishing sites. It advocates for data augmentation, cost-sensitive learning, and domain adaptation to improve the accuracy and generalization of detection models. By highlighting the latest advancements and addressing key challenges, the review contributes to building robust and resilient phishing detection frameworks that safeguard users and organizations in the constantly evolving cyber threat landscape.

A phishing website detection system based on machine learning methods

A phishing website detection system based on machine learning methods

Modelling of an Adaptive Network Model for Phishing Website Detection Using Learning Approaches

Phishing links are spread via text messages, social media platforms, and email by phishing attackers. Social engineering skills are used to visit phishing websites to trick the users and enter critical information related to personal data. The confidential data is stolen to defraud legitimate financial institutions or general websites for illegally attaining the benefits. Many machine learning-based solutions are in the enhancements and the technology of machine learning applications to detect the suggested phishing. The rules are used for a solution which depends on the extracted features, and few features require to lies on the services of third-party that, creating time-consuming and instability in the service of prediction. A deep learning-based framework is suggested to detect website of phishing. A framework is established to determine if there is a risk of phishing in real-time during the web page is visited by the user to give a message of warming by the browser plug-in. The prediction service in real-time merges the various techniques for enhancing the accuracy to lower the fake alarm rates and the time of computation which has the filtering whitelist, interception of the blacklist, and prediction of deep learning (DL). Various models of deep learning are compared using the different datasets in the module of machine learning prediction. The greatest accuracy is obtained as 99.18% by the adaptive Recurrent Neural Networks (a−RNN) model from the results of experiments to demonstrate the suggested feasibility solution.

Comparative Study of Machine Learning Algorithms for Phishing Website Detection

Comparative Study of Machine Learning Algorithms for Phishing Website Detection

Applications of Machine Learning Techniques to Detect Phishing Websites

Phishing attacks are a common way for hackers to obtain sensitive and valuable information from unsuspecting users. These attacks often target critical data such as passwords and financial details. To combat this threat, cybersecurity professionals are constantly searching for reliable and effective techniques for detecting phishing websites. This project investigates the use of machine learning algorithms to identify phishing URLs by extracting and analyzing various features of both legitimate and phishing URLs. The goal is to create a blacklist of known phishing websites that can alert individuals when they browse or access a potentially dangerous site. The project will compare the performance of four machine learning algorithms such as Ensemble Adaboost Classifier, Multi-Layer Perceptron Classifier, Stochastic Gradient Descent classifier, and XGBoost - based on their accuracy, speed, and other factors.

SOLVING THE PROBLEM OF DETECTING PHISHING WEBSITES USING ENSEMBLE LEARNING MODELS

Due to the popularity of the easiest way to obtain personal information among attackers, phishing detection is becoming a popular area for research aimed at countering the implementation of such attacks. Malicious website detection is essential to prevent the spread of malware and protect end users from victims. Unfortunately, malicious URL detection still needs to be better understood due to a lack of features and inaccurate classification. Possible sources were examined in order to investigate the subject. Based on the collected information from previous studies, this study is devoted to solving the problem of detecting phishing websites using Ensemble Learning. The aim of the work is to choose the most optimal algorithm for classifying phishing websites using gradient boosting algorithms. AdaBoost, CatBoost, and Gradient Boosting Classifier were chosen as Ensemble Learning algorithms and were used to improve the efficiency of classifiers. Practical studies of the parameters of each algorithm for finding the optimal classification model are given. Research and experiments were carried out on a dataset containing information extracted from the contents of a URL: main URL, domain, directory, and file. A thorough Exploratory Data Analysis (EDA) was carried out, as a result of which the main dependencies and patterns of determining phishing resources were identified using correlation analysis. ROC AUC Score was chosen as an evaluation metric for the algorithms. The best result for predicting phishing websites was demonstrated by the AdaBoost Classifier algorithm, with an average ROC AUC score of 99%. The results of the experiments were illustrated in the form of graphs and tables.

Phishing Website Detection Using Several Machine Learning Algorithms: A Review Paper

Phishing is one of the major web social engineering attacks. This has led to demand for a better way to predict and stop them in a commercial environment. This paper seeks to understand the research done in the field and analyse the next steps forward. This is done by focusing on what goes into the selection of proper features, from manual selection to the use of Genetic Algorithms such as ADABoost and MultiBoost. Then a look into the classifiers in use, Neural Networks and Ensemble algorithms which were prominent alongside some novel approaches. This information is then processed into a framework for cloud-based and client-based phishing website detection, alongside suggestions for possible future research and experiments that could help progress the field.

Phishing Website Detection as a Website Comparing Problem

Phishing Website Detection as a Website Comparing Problem

Hybrid Approach for Phishing Website Detection Using Classification Algorithms

The internet has significantly altered how we work and interact with one another.Statistics show 63.1 % of the present world population are internet users. This clearly indicates how heavily man is dependent on digital media. Digital media users are on the rise and so is the incidence of cyber crimes. People who lack experience and knowledge are more vulnerable and susceptible to phishing scams.The victims experience severe consequences as their personal credentials are at stake. Phishers use publicly available sources to acquire details about the victim's professional and personal history.Countermeasures must be implemented with the highest priority. Detection of malicious websites can significantly reduce the risk of phishing attempts.In this research, a highly accurate website phishing detection method based on URL features is proposed. We investigated eight existing machine learning classification techniques for this, including extreme gradient boosting (XGBoost), random forest (RF), adaptive boosting (AdaBoost), decision trees (DT), K-nearest neighbors (KNN), support vector machines (SVM), logistic regression and naïve bayes (NB) to detect malicious websites.The results show that XGboost had the best accuracy with a score of 96.71%, followed by random forest and AdaBoost.We further experimented with various hybrid combinations of the top three classifiers and observed that XGboost-Random Forest hybrid algorithms produced the best results.The hybrid model classified the websites as legitimate or phishing with an accuracy of 97.07%.

The Effect of Artificial Intelligence on Service Quality and Customer Satisfaction in Jordanian Banking Sector

The study emphasizes the importance of Artificial Intelligence (AI) and its applications on the service quality provided by Jordanian banks for their customer satisfaction. This research paper thoroughly reviews the literature on the numerous emergent applications of artificial intelligence and its impact on the banking sector. A rigorous study of the available literature is conducted to examine AI's uses in banking. Artificial intelligence improves the banking experience for millions of clients and employees by providing credit score checking, system failure prediction, emergency alarm systems, fraud detection, phishing website detection, liquidity risk assessment, customer loyalty evaluation and intelligence systems by reducing the employee workload. A questionnaire gathered data from 270 consumers in Jordan's banking sector. The SPSS program used exploratory factor analysis to statistically evaluate the sample data to determine service quality and customer satisfaction. The results show that the updated SERVQUAL Model extracted five subscales instead of the eight in the original model. The extracted subscales were tangibility, assurance, reliability, responsiveness, and empathy. According to this study, artificial intelligence is statistically relevant to service quality and customer satisfaction. The updated SERVQUAL model, according to the authors, helps address customer satisfaction in the banking sector. The research findings suggest that the demand for artificial intelligence in the Jordanian banking sector is equally essential for the customers; thus, there should be an optimal balance between virtual and human agents based on the customers' requirements and preferences. Further, this study found practical implications of using AI in banking, particularly those related to Jordanian customer perception.

Vote algorithm based probabilistic model for phishing website detection

Internet scams have been a major concern for everyone over the past decade. With the advancement of technology, attackers have formulated different kinds of contemporary fraudulent procedures to obtain user’s sensitive information. Phishing is one of the oldest and common fraudulent attempts by which every year millions of internet users fall victim to scams resulting in losing their money. Different techniques and algorithms have been proposed by researchers in detecting phishing websites. However, the detection of phishing websites has few challenges since there are different subjective considerations and ambiguities involved in the detection process. This paper presents a two-stage probabilistic method for detecting phishing websites based on the vote algorithm. In the first stage, 29 different base classifiers have been used and their probabilistic values were calculated. In the second stage, the voting algorithm aggregated the probabilistic values of several base classifiers and the phishing websites were detected using the average of probabilities approach. The voting technique achieved an accuracy of 97.431% outperforming all of the single base classifiers in terms of accuracy.<br /><br />

A Proposal Phishing Attack Detection System on Twitter

The security of personal data is crucial for a company or any individual. Phishing is one of the most common and dangerous cybercrime attacks. These attacks aim to steal information used by individuals and organizations using social engineering, which is a key point for the success of the phishing attack. Even though there are several systems and solutions, the amount of personal information stolen continues to increase as cyberattacks become more difficult to detect. This paper consists of a broad review to study the work carried out in the fight against phishing and the identification of vulnerabilities in existing systems to achieve better efficiency. The authors focused on the social medium Twitter to study the phishing attacks passing through this medium, and they present their new design, which is based on new features. The classification of the approach includes 23 features and uses the MLP artificial neural network (ANN MLP) algorithm. Experiments show that the system is effective at detecting phishing sites, with a 96% success rate using recent data.

Homoglyph Attack Detection Model Using Machine Learning and Hash Function

Phishing is still a major security threat in cyberspace. In phishing, attackers steal critical information from victims by presenting a spoofing/fake site that appears to be a visual clone of a legitimate site. Several Unicode characters are visually identical to ASCII characters. This similarity in characters is generally known as homoglyphs. Malicious adversaries utilize homoglyphs in URLs and DNS domains to target organizations. To reduce the risks caused by phishing attacks, effective ways of detecting phishing websites are urgently required. This paper proposes a homoglyph attack detection model that combines a hash function and machine learning. There are two phases to the model approach. The machine was being trained during the development phase. The deployment phase involved deploying the model with a Java interface and testing the outcomes through actual user interaction. The results are more accurate when the URL is hashed, as any little changes to the URL can be recognized. The homoglyph detector can be developed as a stand-alone software that is used as the initial step in requesting a webpage as it enhances browser security and protects websites from phishing attempts. To verify the effectiveness, we compared the proposed model on several criteria to existing phishing detection methods. By using the hash function, the proposed security features increase the overall security of the homoglyph attack detection in terms of accuracy, integrity, and availability. The experiment results showed that the model can detect phishing sites with an accuracy of 99.8% using Random Forest, and the hash function improves the accuracy of homoglyph attack detection.

An improved anti‐phishing model utilizing TF‐IDF and AdaBoost

SummaryA behavior over the internet, which results into false utilization of the resources is termed as phishing and the techniques to identify them are termed as anti‐phishing frames and algorithms. This research work introduces a new relative similarity approach utilizing similarity index co‐relation by applying Cosine similarity. In order to incorporate the co‐relation, term frequency and inverse document frequency (TF‐IDF) is also calculated. AdaBoost is employed to improve the classification accuracy of the proposed work by means of enhancing the performance of weak learner to classify website into legitimate and phishing one. The simulation analysis has shown that the proposed work outperformed the existing works in terms of accuracy, TPR, and precision of detecting phishing sites to offer secure browsing. Overall model evaluation against 50,000 URL instances have demonstrated an average precision, TPR, F‐measure, accuracy and execution time of 98.64%, 96.52%, 97.56%, 98.014%, and 3.485 s, respectively for phishing detection. This proves the success of the proposed model in shielding phishing attack.

Phishing website detection: How effective are deep learning‐based models and hyperparameter optimization?

AbstractPhishing websites are fraudulent websites that appear legitimate and trick unsuspecting users into interacting with them, stealing their valuable information. Because phishing attacks are a leading cause of data breach, different anti‐phishing solutions have been explored for cybersecurity management including machine learning‐based technical approaches. However, there is a gap in understanding how robust deep learning‐based models together with hyperparameter optimization are for phishing website detection. In this vein, this study pursues the tasks of developing parsimonious deep learning models and hyperparameter optimization to achieve high accuracy and reproducible results for phishing website detection. This paper demonstrates a systematic process of building detection models based on three deep learning algorithm architectures (Long Short‐Term Memory‐based detection models, Fully Connected Deep Neural Network‐based detection models, and convolutional neural network‐based detection models) that are built and evaluated using four publicly available phishing website datasets, achieving the best accuracy of 97.37%. We also compared two different optimization algorithms for hyperparameter optimization: Grid Search and Genetic Algorithm, which contributed to 0.1%–1% increase in accuracy.

PHISHING URL DETECTION WEBSITE

Phishing websites are one of the types of internet security issues that target human vulnerabilities rather than software flaws. It's the technique of luring online users in order to get sensitive information like usernames and passwords. We present an intelligent technique for detecting phishing websites in this research. The technology works as an add on to an internet browser, notifying the user when it finds a phishing website. The system is based on supervised learning, which is a type of machine learning. We chose the Random Forest approach because of its high categorization performance. Our goal is to develop a higher performing classifier by examining the characteristics of phishing websites and combining them in the best way possible.

Multi-scale semantic deep fusion models for phishing website detection

In view of semantic counterfeiting characteristics of phishing websites and their multi-scale composition, this paper fully considers the semantic information of different scales, and proposes three semantic-based phishing detection models at different depths using various deep learning methods. The proposed three models are Multi-scale Data-layer Fusion (MDF) model, Multi-scale Feature-layer Fusion (MFF) model and Multi-scale In-depth Fusion(MIF) model. Experimental results on a constructed complex dataset show that the three models all have good recognition capabilities and the MIF model achieves the best performance on a complex dataset, with an F1-Measure of 0.9830, AUC value of 0.9993 and a false positive rate of 0.0047. Then with further comparison with both visual and text methods and an active discovery experiment lasting for 6 months with 3016 phishing websites detected in the real network environment, it is found that the proposed model is both competitive and practical for real detection scenarios.

Evolutionary Algorithm with Deep Auto Encoder Network Based Website Phishing Detection and Classification

Website phishing is a cyberattack that targets online users for stealing their sensitive data containing login credential and banking details. The phishing websites appear very similar to their equivalent legitimate websites for attracting a huge amount of Internet users. The attacker fools the user by offering the masked webpage as legitimate or reliable for retrieving its important information. Presently, anti-phishing approaches necessitate experts to extract phishing site features and utilize third-party services for phishing website detection. These techniques have some drawbacks, as the requirement of experts for extracting phishing features is time consuming. Many solutions for phishing websites attack have been presented, such as blacklist or whitelist, heuristics, and machine learning (ML) based approaches, which face difficulty in accomplishing effectual recognition performance due to the continual improvements of phishing technologies. Therefore, this study presents an optimal deep autoencoder network based website phishing detection and classification (ODAE-WPDC) model. The proposed ODAE-WPDC model applies input data pre-processing at the initial stage to get rid of missing values in the dataset. Then, feature extraction and artificial algae algorithm (AAA) based feature selection (FS) are utilized. The DAE model with the received features carried out the classification process, and the parameter tuning of the DAE technique was performed using the invasive weed optimization (IWO) algorithm to accomplish enhanced performance. The performance validation of the ODAE-WPDC technique was tested using the Phishing URL dataset from the Kaggle repository. The experimental findings confirm the better performance of the ODAE-WPDC model with maximum accuracy of 99.28%.

Phishing websites detection using a novel multipurpose dataset and web technologies features

Phishing attacks are one of the most challenging social engineering cyberattacks due to the large amount of entities involved in online transactions and services. In these attacks, criminals deceive users to hijack their credentials or sensitive data through a login form which replicates the original website and submits the data to a malicious server. Many anti-phishing techniques have been developed in recent years, using different resource such as the URL and HTML code from legitimate index websites and phishing ones. These techniques have some limitations when predicting legitimate login websites, since, usually, no login forms are present in the legitimate class used for training the proposed model. Hence, in this work we present a methodology for phishing website detection in real scenarios, which uses URL, HTML, and web technology features. Since there is not any updated and multipurpose dataset for this task, we crafted the Phishing Index Login Websites Dataset (PILWD), an offline phishing dataset composed of 134,000 verified samples, that offers to researchers a wide variety of data to test and compare their approaches. Since approximately three-quarters of collected phishing samples request the introduction of credentials, we decided to crawl legitimate login websites to match the phishing standpoint. The developed approach is independent of third party services and the method relies on a new set of features used for the very first time in this problem, some of them extracted from the web technologies used by the on each specific website. Experimental results show that phishing websites can be detected with 97.95% accuracy using a LightGBM classifier and the complete set of the 54 features selected, when it was evaluated on PILWD dataset.