رویکرد کارآمد برای ردیابی فیشینگ با استفاده از مدل عصبی فازی

An Efficient Approach for Phishing Detection Using Neuro-Fuzzy Model

Abstract

Nowadays, online transactions are becoming more and more popular in modern society. As a result, Phishing is an attempt by an individual or a group of people to steal personal information such as password, banking account and credit card information, etc. Most of these phishing web pages look similar to the real web pages in terms of website interface and uniform resource locator (URL) address. Many techniques have been proposed to detect phishing websites, such as Blacklist-based technique, Heuristic-based technique, etc. However, the numbers of victims have been increasing due to inefficient protection technique. Neural networks and fuzzy systems can be combined to join its advantages and to cure its individual illness. This paper proposed a new neuro-fuzzy model without using rule sets for phishing detection. Specifically, the proposed technique calculates the value of heuristics from membership functions. Then, the weights are generated by a neural network. The proposed technique is evaluated with the datasets of 11,660 phishing sites and 10,000 legitimate sites. The results show that the proposed technique can detect over 99% phishing sites.

1 Introduction

Phishers use a number of techniques to lure their victims, including email messages, instant messages, forum posts, phone calls, and text messages. With these activities of phishing, it causes severe economy loss all over the world. APWG’s second half report for 2010 claimed that phishing attacks grew 142% over the first half of 2010. The report also classifies the targets as comprising 37.9% payment services, 33.1% financial institutions, 6.6% classified, 4.6% gaming, 2.8% social networks, and the remainder in other categories. In 2011, 83% of Americans and 85% of Europeans regularly shopped online (Fortune Magazine, 2011). Meanwhile, phishing sites are also growing rapidly in quality and quantity. Therefore, the risk of stealing user information is extremely high. Because of these reasons, detecting phishing problem is very urgent, complex and extremely important problem in modern society. Recently, there have been many studies which against phishing based on the characteristics of site, such as URL of website, content of website, combining both the website URL and content, source code of website or screenshot of website, Manuscript received September 21, 2014; revised December 11, 2014. etc. However, each of study has its own strengths and weaknesses. There is still not a sufficient method. In this paper, a new approach is proposed to detect the phishing sites that focuses on the features of URL (PrimaryDomain, SubDomain, PathDomain) and the ranking of site (PageRank, AlexaRank, AlexaReputation). Then, a proposed neuro-fuzzy network is a system which reduces the error and increases the performance. The proposed neuro-fuzzy model uses computational models to perform without rule sets. The proposed solution achieved detection accuracy above 99% with low false signals. The rest of this paper is organized as follows: Section II presents the related works. System design is shown in section III. Section IV evaluates the accuracy of the method. Finally, Section V concludes the paper and figures out the future works.

شناسایی فیشینگ: یک مدل عصبی فازی کارآمد بدون استفاده از مجموعه قوانین

Phishing Identification: An Efficient Neuro-Fuzzy Model Without Using Rule Sets

Abstract

The Internet has brought enormous benefits to mankind, but it could be many potential risks. Internet crimes are growing rapidly, phishing is one of the new type of online crime. Phishing site is a fake-site aimed to steal personal information such as password, banking account and credit card information, etc. Most of these phishing pages look similar to the real pages in terms of interface and uniform resource locator (URL) address. Many techniques have been proposed to identify phishing sites. However, the numbers of victims have been increasing due to inefficient protection technique. In this paper, we develop a neuro-fuzzy model for phishing identification efficiently. The model eliminates the subjective factors to improve efficiency such as if-then rule sets, the parameters of membership functions, etc. Moreover, the efficiency features to identify phishing were used for the neuro-fuzzy model. The effectiveness of the proposed technique is examined with large-scale datasets collected from phishing sites and legitimate sites. The results show that the proposed technique can identify over 99% phishing sites.

1 Introduction

Phisher use a number of techniques to fool their victims aimed to steal personal information including email messages, instant messages, forum post, phone calls and social networking. These activities of phishing cause severe economy loss all over the world. According to Fortune Magazine, in 2011, 83% of Americans and 85% of Europeans regularly shopped online. Meanwhile, phishing sites are also growing rapidly in quality and quantity. Therefore, the risk of stealing user information is extremely high. For of these reasons, identifying phishing problem is very urgent, complex and extremely important problem.

In this paper, an efficient method is proposed to identify the phishing sites that focuses on the features of URL (PrimaryDomain, SubDomain, PathDomain) and Google’s parameters (PageRank, BackLink, GoogleIndex). Then, a proposed neurofuzzy model is a system which reduces the error and increases the performance. The neuro-fuzzy model uses computational models to perform without using if-then rule sets. The proposed technique achieved identification accuracy above 99% with low false signals.

The remainder of this paper is organized as follows: Section II presents the related works. System design is detailed in section III. Section IV evaluates the accuracy of the method. Finally, Section V concludes the paper and figures out the future works.

فیشینگ، اسمیشینگ و ویشینگ: ارزیابی تهدیدات علیه دستگاه های تلفن همراه

Phishing, SMiShing & Vishing: An Assessment of Threats against Mobile Devices

چکیده

1. مقدمه

1.1 بیان مسئله

2. بررسی پیشینه‌

2.1 مهندسی اجتماعی

2.2 فیشینگ، اسمیشینگ و ویشینگ

2.3 دستگاه‌های تلفن همراه

2.4 تهدیدها علیه دستگاه‌های همراه

2.5 داده کاوی متن

3. روش تحقیق

3.1 نشانه‌های کلی

3.2 نشانه‌های فیشینگ

3.3 نشانه‌های اسمیشینگ

4. یافته‌ها

4.1 اطلاعات کاربر از امکانت آنلاین

4.2 رفتار آنلاین کاربر

4.3 اطلاع کاربر از تهدیدهای آنلاین

5. نتیجه‌گیری

چکیده

این تحقیق یک ارزیابی اکتشافی حملات فیشینگ، اسمیشینگ و ویشینگ علیه دستگاه‌های تلفن همراه است. این پژوهش به بررسی پیامد‌های رفتار کاربر نهایی جهت کاهش تهدید‌های ناشی از استفاده‌ی دستگاه‌های تلفن همراه برای خدمات و امکانات آنلاین پرداخته است. فیشینگ عبارت است از فریب کاربر از طریق حمله‌ی اجتماعی مهندسی شده با هدف کلاهبرداری از کاربران خوش باور با وب‌سایت‌های آشنا که جعلی بوده و طوری به نظر می‌رسد که از یک سازمان و منبع قانونی است. این کاربر را فریب می‌دهد تا مهاجم را مجهز به اعتبارنامه‌های دسترسی خود کند تا دسترسی مجاز برای زیان زدن به کاربر حاصل شود. حملات اسمیشینگ زمانی اتفاق می‌افتد که پیام متنی برای کاربر ارسال می‌شود یا بر لینک آماده شده کلیک می‌کند، که منجر به یک وب‌سایت جعلی یا دسترسی مهاجم به مخاطبان کاربر و/یا هر اطلاعات اعتباری دیگری می‌شود. ویشینگ یک حمله‌ی فیشینگ صوتی است که یک تماس صوتی از سوی یک مهاجم دریافت می‌شود و هدف را فریب داده تا اطلاعات شخصی‌اش را در اختیار مهاجم قرار دهد که هدف مهاجم استفاده از آن اطلاعات برای زیان زدن به کاربر است. امروزه با گسترش و تکثیر تلفن‌های هوشمند، تبلت‌ها و هات‌اسپات‌ها(وای-فای) این حملات اجتماعی مهندسی شده به دستگاه‌های تلفن همراه رایج هستند. این پژوهش، به مشاهده‌ی 20 کاربر نهایی از لحاظ آگاهی، درک و رفتار آنها در مواقع مواجه با موقعیت‌های حمله‌ی فیشینگ پرداخته و با آنها در این زمینه مصاحبه انجام داده است. نتایج نشان داد که مردان خیلی راحت‌تر به فضای سایبری اعتماد می‌کنند و بنابراین، بیشتر از زنان در معرض حملات فیشینگ هستند. همچنین نتایج نشان داد، بیشتر کاربران یا تا حد کمی از تهدید‌های فیشینگ، اسمیشینگ و ویشینگ علیه دستگاه‌های تلفن همراهشان آگاهند یا اصلا آگاهی در این مورد ندارند. جالب توجه است که 55% گاهی اوقات پیام‌های دریافتی به صورت تهدید‌ها را امتحان می‌کنند در‌حالیکه، 35% هرگز هیچ پیامی از این نوع را بررسی نمی‌کنند. یک طبقه‌بندی از کلمات به‌”دام انداختن”(alluring)و “فریب دادن”(decoying)در حملات فیشینگاستفاده شده است که به عنوان معیاری برای کاربران نهایی جهت محافظت علیه قربانی سایبری شدن،تدارک دیده شده است. از میان رایج‌ترین سیستم‌های عاملiOs آسیب‌پذیر‌ترین نوع در مقابل حملات فیشینگ می‌باشد.

5. نتیجه‌گیری

در این بخش نتیجه‌گیری یافته‌های اصلی خلاصه شده، محدودیت‌های تحقیق مطرح شده و ملاحظات پایانی برای تحقیقات بیشتر مطرح شده‌اند.

هدف این پژوهش، مشخص نمودن تهدیدهای مختلفی بود که علیه دستگاه‌های همراه در مبارزه‌اند و اطلاع کاربران نهایی از این تهدیدها. ما تلاش کردیم به این بپردازیم که تا چه میزان حملات فیشینگ، دستگاه‌های همراه را تحت تاثیر قرار می‌دهند. بطور کلی، ملاحظه شد که مردان دانش و معلومات فناوری کافی در مورد عملکردهای خدمات و امکانات اینترنت دارند. به همین نحو، مشخص شد که آنها در فضای سایبری بسیار احساس راحتی کرده و به آن اعتماد دارند، بنابراین، همین امر آنها را نسبت به زنان هم‌کیش خود بیشتر در معرض حملات تلفن همراه قرار می‌دهد.

Abstract

This study is an exploratory assessment of Phishing, SMiShing and Vishing attacks against mobile devices. It examines the implications of end-user behavior towards mitigating the risks posed by using mobile devices for online services and facilities. Phishing is that socially engineered attack aimed at enticing unsuspecting users with familiar websites spoofed and purported to come from a legitimate organization or source. It lures the user to furnish the assailant with the user’s access credentials, for which privileged access would be gained to harm the user. SMiShing attacks also happen whenever text messages are sent for the user to either click on a link provided, which leads to a fraudulent website or for the attacker to get access to the user’s contacts and/or any other confidential information. Vishing is a voice phishing attack, whereby a voice call received from an assailant lures the target into providing personal information with the intention to use that information to cause harm. With the proliferation of smart phones, tablets and hotspots, these social engineering attacks on mobile devices are now prevalent. The study observed and strategically interviewed 20 end-users for their knowledge, perceptions and behavior when confronted with phishing attack situations. The results show that men are more comfortable and trusting on the cyber-space and thus more susceptible to phishing attacks than women. The results also indicate that most users are either slightly aware or not at all aware of Phishing, SMiShing and Vishing threats against their mobile devices. Interestingly, 55% would occasionally examine the messages received as perceived threats, whilst 35% would never or almost never scrutinize any messages. A taxonomy of ‘alluring” and “decoying” words used in phishing attacks is provided as a benchmark to end-users to guard against becoming cyber-victims. Of the most commonly used operating systems, the iOS was found to be the most susceptible to phishing attacks.

5. CONCLUSION

This concluding section summarizes the key findings, discusses the study’s limitations and makes concluding remarks for future research.

The object of this study was to identify the various threats that militate against mobile devices and the behavior and perceptions of end-users towards those threats. We endeavored to address the extent to which phishing attacks affect mobile devices. Overall, men were perceived to have adequate technological know-how of the operations of the Internet services and facilities. Similarly, they were found to be so comfortable and trusting whenever on the cyber-space, thus making them more susceptible to mobile attacks than their women folks.

تصویری از مقاله ترجمه و تایپ شده در نرم افزار ورد

فیشینگ، اسمیشینگ و ویشینگ: ارزیابی تهدیدات علیه دستگاه های تلفن همراه

Phishing, SMiShing & Vishing: An Assessment of Threats against Mobile Devices

یک رویکرد عصبی فازی برای شناسایی فیشینگ

A Novel Neuro-Fuzzy Approach for Phishing Identification

Abstract

Together with the growth of Internet, e-commerce transactions play an important role in the modern society. As a result, phishing is a deliberate act by an individual or a group of people to steal personal information such as password, banking account, credit card information, etc. Most of these phishing web pages look similar to the real web pages in terms of website interface and uniform resource locator (URL) address. Many techniques have been proposed to identify phishing websites, such as Blacklist-based technique, Heuristic-based technique, etc. However, the number of victims has been increasing due to inefficient protection technique. Neural networks and fuzzy systems can be combined to join its advantages and to cure its individual illness. This paper proposed a new neuro-fuzzy model without using rule sets for phishing identification. Specifically, the proposed technique calculates the value of heuristics from membership functions. Then, the weights are trained by neural network. The proposed technique is evaluated with the datasets of 11,660 phishing sites and 10,000 legitimate sites. The results show that the proposed technique can identify over 99% phishing sites.

1 Introduction

The word ”phishing” is produced from the word ”fishing”. Phishers, creating phishing sites, use a number of techniques to fool their victims, including email messages, instant messages, forum posts, phone calls and social networking. With these activities of phishing, it causes severe economy loss all over the world. According to a study by Gartner [1], 57 million US Internet users have identified the receipt of email linked to phishing scams and about 2 million of them are estimated to have been tricked into giving away sensitive information. Meanwhile, phishing sites are also growing rapidly in quality and quantity. Therefore, the risk of stealing user information is extremely high. Because of these reasons, identifying phishing problem is very urgent, complex and extremely important problem in modern society. Recently, there have been many studies that against phishing based on the characteristics of site, such as URL of website, content of website, combining both the website URL and content, source code of website or interface of website, etc. However, each of studies has its own strengths and weaknesses. There is still not a sufficient method. In this paper, a new approach is proposed to identify the phishing sites that focuses on the features of URL (PrimaryDomain, SubDomain, PathDomain) and the ranking of site (PageRank, AlexaRank, AlexaReputation. Then, a proposed neuro-fuzzy network is a system which reduces the error and increases the performance. The proposed neuro-fuzzy model uses computational models to perform without rule sets. The proposed solution achieved identification accuracy above 99% with low false signals.

The rest of this paper is organized as follows: Section II presents the related works. System design is shown in section III. Section IV evaluates the accuracy of the method. Finally, Section V concludes the paper and figures out the future works.