Anti-spam filtering: a centroid-based classification approach

Soonthornphisaj, Nuanwan; Chaikulseriwat, K.; Tang-On, Piyanan

doi:10.1109/icosp.2002.1179980

Cited by 13 publications

(9 citation statements)

References 1 publication

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The statistical method does not depend on a particular language context, but only has an ability of identifying spam mail and legitimate mail [11][12][13]. It is difficult to classify using the statistical method.…”

Section: An Intelligent Algorithm Based On Vector Spacementioning

confidence: 99%

A Composite Intelligent Method for Spam Filtering

Liu¹,

Chen

Liu³

et al. 2014

IJSIA

View full text Add to dashboard Cite

This paper analyses several common algorithms for spam filtering and shows the advantages and disadvantages of these algorithms for spam filtering. Each algorithm is only suitable for filtering specific spam. Some algorithms are suitable for Chinese, and some algorithms perform well in English. In a lot of spam, it is not reliable and inefficiency to using a single algorithm to separate out spam. Thereby, in order to improve the accuracy and efficiency of spam filtering, composite intelligent algorithm, which integrates and improves the existing algorithms by utilizing the advantages of previous algorithms and avoiding their shortages, is proposed. Moreover, an intelligent method that it has the ability of self-learning by using the contents of the e-mails is introduced. Finally, the outcome of experiment shows that the intelligent method achieves a better efficiency and performance.

show abstract

Section: An Intelligent Algorithm Based On Vector Spacementioning

confidence: 99%

A Composite Intelligent Method for Spam Filtering

Liu¹,

Chen

Liu³

et al. 2014

IJSIA

View full text Add to dashboard Cite

show abstract

“…Literature on spam filtering exists for email [Sahami et al 1998;Salton and McGill 1998;Mitchell 1997;Soonthornphisaj et al 2002;Sakkis et al 2003;Cohen 1996;Rigoutsos and Huynh 2004], and VoIP [Rosenberg et al 2006, Shin et al 2005 infrastructures. But, neither the email nor the VoIP antispam solutions advocate the computation of nuisance due to unwanted messages or calls by using end user's behavior and context.…”

Section: Introductionmentioning

confidence: 98%

Nuisance level of a voice call

Kolan

Dantu

Cangussu

2008

ACM Trans. Multimedia Comput. Commun. Appl.

View full text Add to dashboard Cite

In our everyday life, we communicate with many people such as family, friends, neighbors, and colleagues. We communicate with them using different communication media such as email, telephone calls, and face-to-face interactions. While email is not realtime and face-to-face communications require geographic proximity, voice and video communications are preferred over other modes of communication. However, real-time voice/video calls may create nuisance to the receiver. In this article, we describe a mathematical model for computing nuisance level of incoming voice/video calls. We computed the closeness and nuisance level using the calling patterns between the caller and the callee. To validate the nuisance model, we collected cell phone call records of real-life people at our university and computed the nuisance value for all voice calls. We validated the nuisance levels using the feedback from those real-life people. Such a nuisance model is useful for predicting unwanted voice and video sessions in an IP communication network.

show abstract

“…content of the email and label spam email as BULK and Some spammers are telemarketers, who broadcast advertise-expect the recipients to make a decision on the authenticity ments to thousands or millions of email users. They do not of the source [1], [2], [3], [4], [5], [6], [7], [8], [10], [9], have a specific target. Opt-in spammers constitute another kind [11].…”

Section: Introductionmentioning

confidence: 99%

Unwanted SMTP Paths and Relays

Palla

Dantu

2007

2007 2nd International Conference on Communication Systems Software and Middleware

View full text Add to dashboard Cite

Based on the social interactions of an email user, category. Phishing attackers masquerade the identity of a incoming email traffic can be divided into different categories genuine sender and steal consumers' personal identity data such as, telemarketing, Opt-in family members and friends. Due and financial account credentials. Phishers send spoofed emails to a lack of knowledge in the different categories, most of the existing spam filters are prone to high false positives and false and lead consumers to counterfeit websites designed to trick negatives. Moreover, a majority of the spammers obfuscate their the recipients into divulging their financial information such email content inorder to circumvent the content-based spam as, credit card numbers, account user names, passwords, and filters. However, they do not have access to all the fields in the social security numbers. Legislation cannot help because these email header. Our classification method is based on the path attacks can originate outside the United States.traversed by email (instead of content analysis) since we believe that spammers cannot forge all the fields in the email header. WeWe need a solution for unsolicited emails. The solution based our classification on three kinds of analyses on the header: should consider other aspects of the email such as, information i) EndToEnd path analysis, which tries to establish the legitimacy about how an email is routed from a source to the destination of the path taken by an email and classifies them as either spam besides considering the content of an email. In essence, the or non-spam; ii) Relay analysis, which verifies the trustworthiness reputation and trustworthiness of the relays and the path of the relays participating in the relaying of emails; iii) Emails traversed by an email have to be taken into consideration. wantedness analysis, which measure the recipients wantedness of the senders emails. We use the IMAP message status flags The information about the paths and relays can be obtained such as, message has been read, deleted, answered, flagged, and from the email header. Spammers can obfuscate the content draft as an implicit feed back from the user in Emails wantedness of an email but they cannot manipulate all the fields of the analysis. Finally we classify the incoming emails as i) socially close email header. Our classification method is based on the path (such as, legitimate emails from family, and friends), ii) socially distinct emails from strangers, iii) spam emails (for example, traversed by an email. Though spammers can forge certain emails from telemarketers, and spammers) and iv) opt-in emails. fields of the email header (for example, spammers can insert Based on the relation between spamminess of the path taken by spurious Received: header lines before dispatching the spam spam emails and the unwantedness values of the spammers, we email), they cannot modify the complete path traversed by classify spammers as i) prospective spammers, ii) suspects, iii) an email. Hence we rely on th...

show abstract

Anti-spam filtering: a centroid-based classification approach

Cited by 13 publications

References 1 publication

A Composite Intelligent Method for Spam Filtering

A Composite Intelligent Method for Spam Filtering

Nuisance level of a voice call

Unwanted SMTP Paths and Relays

Contact Info

Product

Resources

About