Evolving Fuzzy Neural Network for Phishing Emails Detection

International Journal of Innovative Technology and Exploring Engineering, 2019

The objective of this undertaking is to apply neural systems to phishing email recognition and assess the adequacy of this methodology. We structure the list of capabilities, process the phishing dataset, and execute the Neural Network frameworks. we analyze its exhibition against that of other real Artificial Intelligence Techniques – DT , K-nearest , NB and SVM machine.. The equivalent dataset and list of capabilities are utilized in the correlation. From the factual examination, we infer that Neural Networks with a proper number of concealed units can accomplish acceptable precision notwithstanding when the preparation models are rare. Additionally, our element determination is compelling in catching the qualities of phishing messages, as most AI calculations can yield sensible outcomes with it.

2020

AbstractPhishing is a crime that is portrayed as a craft of cloning a site page of a real organization with the intent of getting private information of clueless web users. With the assistance of Machine learning algorithms like Random Forest, Decision Tree, Neural system and Linear model we can classify information into phishing, suspicious and genuine. This should be possible dependent on extraordinary features of phishing sites and client doesn't have to check singular sites. Or maybe we can distinguish and anticipate phishing, suspicious and authentic sites by removing some uncommon features. The intent of this work was to create model to protect clients from the phishing assault. Recent researches indicate that a number of phishing detection algorithms have been introduced into the cyber space, however, most of them depend on an existing blacklist or white list for classification. Hence, when another phishing page is presented, the detection algorithms find it hard to effec...

Romanian Cyber Security Journal

The rapid evolution of phishing attacks poses a significant threat to email security, with traditional detection methods struggling to keep up with the constantly changing landscape. This study investigates the use of Artificial Intelligence (AI) for developing adaptive phishing detection techniques capable of identifying and preventing sophisticated phishing attempts. By harnessing the power of AI, the proposed approach aims to enhance email security and protect users from the evolving tactics employed by attackers. This paper explores various machine learning algorithms, including deep learning and natural language processing techniques, to create an intelligent system capable of detecting phishing emails in real time. This system is designed to learn from a diverse dataset of phishing and legitimate emails, dynamically updating its detection capabilities in response to emerging trends and attack patterns. Additionally, the study examines the role of AI in identifying social engineering tactics, such as the use of personalized content and psychological manipulation, which are frequently employed in targeted phishing campaigns.

— With the enlargement of contemporary technologies and the large-scale global computer networks web-attacks are escalating because of emergent curiosity of people and lawful institutions towards internet. Phishing is one of web-attack carried out by attacker using both social and technical engineering. Generally on web more attacks are launched every month with seek of crafting web addict to consider that they are contacting with a legalized entity for the intention of embezzle identity information, logon records and account details. The phishing attack detection and classification methods are utilized for the prevention and in-depth analysis of the attacks. In this paper, the proposed model has been designed with the multi-directional feature analysis along with the Back-Propagation Probabilistic neural network (BP-PNN) classification. The proposed model has performed better in the terms of the accuracy in all of the domains based upon the attack detection and classification.

International Journal of Recent Technology and Engineering

Phishing email becomes more dangers problem in online bank truncation processing problem as well as social networking sites like Facebook, twitter, Instagram. Normally phishing is carrying out by mocking of email or text embedded in email body, which will provoke users to enter their credential. Training on phishing approach is not so much effective because users are not permanently remember their training tricks, warning messages.it is totally depend on the user action which will be performed on certain time on warning messages given by software while operating any URL. In this paper, phishing email classification is enhanced using J48, Naïve Bayes and decision tree on Spam base dataset. J48 does best classification on spam base which is 97%for true positive and 0.025% false negative. Random forest work best on small dataset that is up to 5000 and number of feature are 34.but increase dataset size and reduce feature Naïve Bayes work faster.

Phishing attacks are one of the trending cyber-attacks that apply socially engineered messages that are communicated to people from professional hackers aiming at fooling users to reveal their sensitive information, the most popular communication channel to those messages is through users’ emails. This paper presents an intelligent classification model for detecting phishing emails using knowledge discovery, data mining and text processing techniques. This paper introduces the concept of phishing terms weighting which evaluates the weight of phishing terms in each email. The pre-processing phase is enhanced by applying text stemming and WordNet ontology to enrich the model with word synonyms. The model applied the knowledge discovery procedures using five popular classification algorithms and achieved a notable enhancement in classification accuracy; 99.1% accuracy was achieved using the Random Forest algorithm and 98.4% using J48, which is –to our knowledge- the highest accuracy rate for an accredited data set. This paper also presents a comparative study with similar proposed classification techniques

Phishing email is one of the major issues of the web nowadays ensuing in monetary losses for organizations and individual users. Varied approaches are developed to filter phishing emails. The current paper focuses on machine learning applications used to detect and predict phishing emails. Moreover, a comparative study and analysis of the assorted filtering ways on the market for organizations and individual users is additionally undertaken. The conducted survey is an organized guide to support the current state of the literature, in read of the subject's wide-ranging scope of papers.

2017 Computing Conference, 2017

A phishing email is a legitimate-looking email which is designed to fool the recipient into believing that it is a genuine email, and either reveals sensitive information or downloads malicious software through clicking on malicious links contained in the body of the email. Given that phishing emails cost UK consumers £174m in 2015, this paper proposal is driven by a problem whose resolution will have a great impact on people's lives in the UK and in the world. In this paper, we proposed a Neural Network (NN)-based model for detections and classifications of phishing emails using publically available email datasets for both benign and phishing emails. The results of the experiments are presented in order to demonstrate the effectiveness of the model in terms of accuracy, true-positive rate, false-positive rate, network performance and error histogram.

ISAR Journal of Science and Technology, 2023

The most extensive threat type, for the present moment at least, remains phishing attacks that rely on people's susceptibility to trick them into sharing personal data. Such attacks generally consist of fake emails originating from an apparently reliable source, for instance, businesses, banks or government facilities. While normal forms of email filtering prove somewhat useful in the fight against phishing, the techniques are highly unlikely to catch the modern-day complex phishing channels such as zero-day phishing or spear phishing. As a result of this, there is a growing uptake of applying Artificial Intelligence (AI) and Machine Learning (ML) to address email security. They are capable of training on sampled volumes of emails and then using this training to improve the recognition of phishing and non-phishing instances. Here, we propose to work on a system of phishing detection using AI/ ML, which will be instrumentally crucial in making email security reliable and adaptable. The system employs Random Forest, Support Vector Machines (SVM) as well as Neural Networks to classify the emails based on the features extracted from the subject, body as well as links of the emails. By employing both phishing and innocuous email corpus, we trained and tested these models for the purpose of understanding the viability of phishing identification. The achievements of the study were the improved accuracy of detection when compared to conventional approaches and the further reduction of misrecognition, which improves security in general. It should be noted that by integrating a multi-model approach with learning mechanisms, the proposed system is indeed versatile and strong against advanced phishing threats.

International journal of computer applications, 2016

Phishing techniques have not only grown in number, but also in sophistication. Phishers might have a lot of approaches and tactics to conduct a well-designed phishing attack. The on-line banking consumers and payment service providers, those are the main targets of the phishing attacks, are facing substantial financial loss and lack of trust in Internet-based services. In order to overcome these, there is an urgent need to find solutions to combat phishing attacks. Detecting the phishing website is a complex task which requires significant expert knowledge and experience. So far, various solutions have been proposed and developed to address these problems. Most of these approaches are not able to make a decision dynamically on whether the site is in fact phished, giving rise to a large number of false responses. This is mainly due to limitation of the previously proposed approaches, for example depending only on fixed black and white listing database, missing of human intelligence and experts, poor scalability and their timeliness. In this research the application of an intelligent fuzzy-based classification system for e-banking phishing website detection is investigated and developed. The main aim of the proposed system is to provide protection to the users from phisher"s deception tricks, giving them the ability to detect the legitimacy of the websites. The proposed intelligent phishing detection system employed Fuzzy Logic (FL) model with association classification mining algorithms. The approach combined the capabilities of fuzzy reasoning in measuring imprecise and dynamic phishing features, with the capability to classify the phishing fuzzy rules.