Abstarct With the growth of World Wide Web and large number Hosts are join continuously to the internet, huge number of access events to Web sites pages were recorded by Servers in log files , many users share, send, post and download lot of things from Web Sites, this manner can be difficult to many organization and Agents in order to monitor and control that, the recorded information and type of analysis used to extract useful knowldege and understanding it become a practical challenges to many researchers. Log files can provided many events information regard to Clients activities, server activities and so on. Many organization employee many log files analysis tools to predict, analysis and monitor users behavior towards site contents .In this paper we proposed algorithms to analysis hidden information contents in Log files and discovering patterns by identified users along with them navigation behaviors then clustering similar users based on different interesting log file content for many Web sites that hosted in Web server. Find statistics for every part in log file command line which are not present in many log files analysis tools are supported here and finally discovering frequent Web sites-Users and user s activities towards those Web sites
Keywords: Web Usage Mining ; pattern mining; Clustering; Log file analysis; Web log data.
INTRODUCTION Information on internet and especially on Web sites increasing rapidly day by day, Web Sites play an important role in this manner where authenticated users, users are always uploads, downloads, browsed many contents according to them needs and interest. Web Server provide a way to browse Web Sites by assigning an IP address or DNS to identify it in addition hosted in it , Server record every events in the form of log file. The process of discovering hidden information from Web log file is called Web Mining. The aim of it is to obtain information about navigational behavior and retrieve useful information from very large raw data, can be represented by several millions of event records in log file. Web log data contains different kinds of information and including web document, web structure and user profiles. Web mining classified into three categories depend on which part of Web to be mined [1, 2]. Categories are Web Structure Mining, Web Content Mining and Web Usage Mining, Figure 1 illustrate the categories of Web Mining algorithms.
Web Structure Mining is the task for discovering knowledge from the structure of hyperlinks within Web pages and given useful information for the relationship among Web pages [3, 4]. The clustering process can play important role here by grouping the Web pages based on their structure, pages can represented by nodes and their links as edges among these nodes, clustering process can be done here based on graph representation and understanding structure of Web pages and its related to other pages in other Web Site Pages . Link structure in Web pages can be classified into two types: First, hyperlinks that connect different parts in the same page (Intra). Second, hyperlinks that connect two or more different pages (Inter). The other role can be applied here by identify trustworthy pages and their hub pages for a given subject. Trustworthy pages contain important information and supported by several links referred to it that means these pages are highly referenced. Hub pages contain many links to trustworthy pages that can give a role for clustering Pages based on trustworthy pages. Web Structure Mining can be employed to efficiently improve information retrieval and document classification tasks [5]. Web Content Mining is the task of discovering different kinds of information contents and improving efficient mechanisms to organize and grouping (clustering) multimedia content to the search engines for accessing these contents by using keywords, categories, related contents etc. Multimedia contents on Web pages are varied such as structured content (i.e. XML documents), Semi-structured (i.e. HTML pages), Unstructured content (i.e. plaint text), other related contents Images, Audios, Videos which are added to those pages or linked to other hosted Sites. Recently there are some challenges appear regard to that in the case of many Web sites were designed by using not only HTML language , other Languages and systems were invited here such as Content Management Systems ( CMS) etc. and the plait texts here are encrypted and stored in an SQL data bases and users events were recorded as visited articles and in this case need to combine web mining algorithms in case to mining clustering and extracted useful information from user behaviors and contents related. Web CMS is responsible for storing, control and management data and other component in long-term uses. CMS consist of repository used to store and preserve various component and use various databases to store it. Repository in CMS contain two categories, the first one comprises source files as well as CMS configuration files , these files contain information about type of content, metadata, users and group of users along with them access data , profiles and preferences. The second repository contain databases where content and files will be processed through CMS and inherit