Report : 鄭鄭鄭 Advisor: Hsing-Kuo Pao 1 Learning to Detect Phishing Emails I. Fette, N. Sadeh, and A. Tomasic. Learning to detect phishing emails. In Proceedings of the International World Wide Web Conference (WWW), pages 649–656, 2007.
Jan 08, 2016
Report : 鄭志欣Advisor: Hsing-Kuo Pao
1
Learning to Detect Phishing Emails
I. Fette, N. Sadeh, and A. Tomasic. Learning to detect phishing emails. In Proceedings of the International World Wide Web Conference (WWW), pages 649–656, 2007.
Outline
2
Introduction MethodEmpirical evaluationConclusion
Introduction
3
Phishing (Spoofed websites)Stealing account informationLogon credentialsIdentity information
Phishing Problem – Hard
Method
4
PILFER – A Machine Learning based approach to classification.phishing emails / ham (good) emailsFeature Set
Features as used in email classification
Features as used in email classification
5
IP-based URLs:http://192.168.0.1/paypal.cgi?fix_account Phishing attacks are hosted off of
compromised PCs. This feature is binary.
6
Age of linked-to domain namesLegitimate-sounding domain name
Palypal.com paypal-update.com
These domains often have a limited life WHOIS query
date is within 60 days of the date the email was sent – “fresh” domain.
This is a binary feature
7
Nonmatching URLsThis is a case of a link that says paypal.com
but actually links to badsite.com.
Such a link looks like <a href="badsite.com"> paypal.com</a>.
This is a binary feature.
8
“Here” links to non-modal domain“Click here to restore your account access”
Link with the text “link”, “click”, or “here” that links to a domain other than this “modal domain”
This is a binary feature.
9
HTML emailsEmails are sent as either plain text, HTML, or
a combination of the two - multipart/alternative format.
To launch an attack without using HTML is difficult.
This is a binary feature.
10
Number of linksThe number of links present in an email.
<a> in HTML tag
This is a continuous feature.
11
Number of domainsSimply take the domain names previously
extracted from all of the links, and simply count the number of distinct domains.
Look at the “main” part of a domain https://www.cs.university.edu/ http://www.company.co.jp/
This is a continuous feature.
12
Number of dotsSubdomains like
http://www.my-bank.update.data.com.Redirection script, such as
http://www.google.com/url?q=http://www.badsite.com
This feature is simply the maximum number of dots (`.') contained in any of the links present in the email, and is a continuous feature.
13
Contains javascriptAttackers can use JavaScript to hide
information from the user, and potentially launch sophisticated attacks.
An email is flagged with the “contains javascript” feature if the string “javascript” appears in the email, regardless of whether it is actually in a <script> or <a> tag
This is a binary feature.
14
Spam-filter outputThis is a binary feature, using the trained
version of SpamAssassin with the default rule weights and threshold.
“Ham” or “Spam”This is a Binary feature.
Empirical Evaluation
15
Machine-Learning Implementation Testing Spam Assassin Datasets Additional ChallengesFalse Positives vs. False Negatives
16
Machine-Learning Implementation-PILFERFirst, run a set of scripts to extract all the
features listed.Second , we train and test a classifier using
10-fold cross validation. Random Forest (classifier)
Random forests create a number of decision trees and each decision tree is made by randomly choosing an attribute to split on at each level, and then pruning the tree.
17
• we use a random forest as a classifier.
18
Testing SpamAssassinSpamAssassin is a widely-deployed freely-
available spam filter that is highly accurate in classifying spam emails.
We classify the exact same dataset using SpamAssassin version 3.1.0, using the default thresholds and rules.
Using “Untrain” SpamAssassin “Training” on 10-fold
19
DatasetsTwo publicly available datasets.
ham corpora from the SpamAssassin project6950 non-phishing non-spam emails
Phishingcorpusapproximately 860 email messages
20
Additional ChallengesThe age of the dataset. Phishing websites are short-lived. Some of our features can therefore not be
extracted from older emails, making our tests difficult. EX: Domain linked to
Result
21
22
Conclusion
23
it is possible to detect phishing emails with high accuracy by using a specialized filter, using features that are more directly applicable to phishing emails than those employed by general purpose spam filters.
Reference
24
I. Fette, N. Sadeh, and A. Tomasic. Learning to detect phishing emails. In Proceedings of the International World Wide Web Conference (WWW), pages 649–656, 2007.
www.ics.uci.edu/.../Learning%20to%20Detect%20Phishing%20Emails.pptx
http://armorize-cht.blogspot.com/2010/01/phishing-mail.html
25