期刊名称：Journal of Theoretical and Applied Information Technology
印刷版ISSN：1992-8645
电子版ISSN：1817-3195
出版年度：2017
卷号：95
期号：17
页码：4103
出版社：Journal of Theoretical and Applied
摘要：Clustering email content is a way of categorization of email content based on specific criteria to obtain the category types of emails that can be used for content profiling email. There are several types and content of the email on a mail server that is not desired by the client so that if this occurs continuously will result in performance mail server will be disrupted. Email content profiling was conducted to determine patterns and behavior in the Postfix SMTP email delivery server using the K-means algorithm to know the type of email that can disrupt the performance of the mail server. The profiling process begins with taking the log file content stored on the Postfix Simple Mail Transfer Protocol (SMTP) email server and then analyze the log file using Clustering techniques with K-means algorithm for Profiling email content on Postfix-based mail server using three criteria, namely: message_id, access from and content. Email content analysis was conducted using K-Means Clustering until the 3rd iteration. The clustering results obtained from four categories of e-mail stored in the postfix mail server, that is a true email, fraud, advertising, and emails which do not have a clear purpose. The step profiling using the deductive method to the content of the email that will be obtained information about the type and characteristics of e-mail stored in the Postfix SMTP server. Using the deductive method for the profiling is the solution to assign profiles based on data obtained from the scene, in this case, is the content of the email on the SMTP Mail Server. There are 3 main things done in this deductive technique, namely: analysis of the operation mode refers to the clustering process, as a characteristic signature in email delivery and victimology identified as a victim in email delivery.
