Indian Journal of Artificial Intelligence and Neural Networking (IJAINN) ISSN: 2582-7626 (Online), Volume-1 Issue-2, April 2021 18 Published By: Lattice Science Publication © Copyright: All rights reserved. Retrieval Number: 100.1/ijainn.B1019021221 DOI:10.35940/ijainn.B1019.041221 Journal Website: www.ijainn.latticescipub.com Abstract: Phishing causes many problems in business industry. The electronic commerce and electronic banking such as mobile banking involves a number of online transaction. In such online transactions, we have to discriminate features related to legitimate and phishing websites in order to ensure security of the online transaction. In this study, we have collected data form phish tank public data repository and proposed K-Nearest Neighbors (KNN) based model for phishing attack detection. The proposed model detects phishing attack through URL classification. The performance of the proposed model is tested empirically and result is analyzed. Experimental result on test set reveals that the model is efficient on phishing attack detection. Furthermore, the K value that gives better accuracy is determined to achieve better performance on phishing attack detection. Overall, the average accuracy of the proposed model is 85.08%. Keywords: Phishing attack, Machine learning, KNN Network security, Phishing detection. I. INTRODUCTION Phishing is a type of social engineering attack where the attacker attempts to gain access to a system by collecting sensitive information such as user name, password and credit card details [1]. The attacker uses forged websites to collect the sensitive information from online users such as mobile bank customers. Once, the sensitive information is gathered though forged websites, then the attacker gets access to a system and such access causes financial loss to the legitimate users. Numerous methods and countermeasures have been proposed to safeguard users from phishing attack [2]. One of phishing attack detection approaches is content-based anti-phishing. Content-based anti-phishing approach is visual similarity employed to identify the contents of phishing websites form legitimate website by analyzing the similarity of contents. Another approach to phishing attack detection is classification of URL into malicious and legitimate classes by employing machine-learning algorithm and proposing model that automatically detects malicious URL and takes an action to stop access to such URLs. Although, different approaches have been proposed for detecting phishing attack, phishing attack remained a major Manuscript received on March 31, 2021. Revised Manuscript received on April 05, 2021. Manuscript published on April 10, 2021. * Correspondence Author Tsehay Admassu Assegie*, Department of Computer Science, Faculty of Computing Technology, Aksum Institute of Technology, Aksum University, Axum, Ethiopia. Email: tsehayadmassu2006@gmail.com © The Authors. Published by Lattice Science Publication (LSP). This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/ ) challenge to the business industry involving online transaction. One of the major challenges of phishing attack detection is that attackers are adapted to different phishing attack detection approaches. Consequently, an effective and efficient phishing detection approach is important to tackle the problem of phishing attack. Hence, we are motivated in designing and implementing an alternative approach to detect phishing attack more efficiently with machine learning. Overall, the contribution of this study is to provide an effective model to phishing attack detection by proposing machine-learning model that classifies URL into phishing or legitimate classes automatically. Therefore, the objectives of this study is to explore the answers to the following questions:  How to create a classification model by employing KNN algorithm to detect phishing attack?  What is the accuracy of KNN algorithm on phishing attack detection?  What is the value of K that gives optimal accuracy on phishing attack detection? II. LITERATURE REVIEW Numerous studies has been conducted on phishing attack detection problem and a number of approaches have been proposed although phishing attack detection still remains a major challenge and much work is required to overcome the challenges of phishing attack. This section presents a review of recently published studies related to phishing attack detection using automated predictive model for decision support on identifying whether a given URL is suspicious or legitimate.In [3], the authors proposed decision tree based model for phishing attack detection. The authors employed 11,055 observation of suspicious or phishing and nonsuspicious or legitimate websites with 30 features. The performance of the proposed model is evaluated on test set and result shows that decision tree algorithm is effective on phishing attack detection, although the accuracy is promising, there is still larger scope for improving the performance to get better results.In another study [4], conducted on phishing attack detection problem, random forest algorithm is employed to phish tank dataset and model for phishing attack detection is proposed to classify RUL into malicious or phishing and legitimate classes. The performance of the proposed is evaluated and the result shows that the proposed model has acceptable accuracy on phishing attack detection. In [5], convolutional neural network based intelligent phishing attack detection model is proposed. K-Nearest Neighbor Based URL Identification Model for Phishing Attack Detection Tsehay Admassu Assegie