Re: Dealing with Web Mining on Apache Log

From: Peter Kim [MS] (peterkim_at_online.microsoft.com)
Date: 03/02/04 

Date: Mon, 1 Mar 2004 17:57:16 -0800

The first question is data preparation issue that the MS Analysis Services
DM component doesn't directly addresses. I'm forwarding to DTS group in case
they have a suggestion.

Once you load, the first two questions could be answered by simple SQL
queries. The other two are more sequence analysis problems. MS Analysis
Services 2000 DM component doesn't directly support sequence analysis, but
you could use decision trees and clustering algorithms to analyze the log
without ordering being modelled. I believe DBMiner(www.dbminer.com) has an
implementation for sequence analysis as an aggregated provider of MS
Analysis Services. 

-- 
Peter Kim
This posting is provided "AS IS" with no warranties, and confers no rights.
"jenny" <anonymous@discussions.microsoft.com> wrote in message
news:48cc01c3fedb$91cd2720$a101280a@phx.gbl...
> I am new to Data Mining but I need to develop a tool for
> mining the apache log file for my final year project.
> I have several questions to ask.
>
> 1. How can I insert apache log file like access_log and
> referer_log into the sql server 2000, any tools can load
> all
> the data immediately?
>
> Format of the access_log:
> 218.102.21.133 - - [01/Sep/2003:00:00:03
> +0800] "GET /cslab/pics/d_hours.gif HTTP/1.1" 304 -
>  "http://www.cs.cityu.edu.hk/cslab/left.html" "Mozilla/4.0
>  (compatible; MSIE 6.0; Windows NT 5.0)"
>
> Format of the referer_ log:
> http://www.cs.cityu.edu.hk/~fypms/student_menu.cgi?
> show_proposal.cgi -> /~fypms/student.cgi
>
> 2. Is it possible to get statistics result like
> a. the greatest number of hits
> b. where the potential applicants came from
> c. where they will go after visiting the main page
> d. show the users' access pattern over a certain period
> time (equals to revisit the same pages)
>
> 3. Can ASP handle the above work with sql server 2000?
>
> Would anyone give me a hand on it?

Relevant Pages

  • Re: Dealing with Web Mining on Apache Log
    ... The first question is data preparation issue that the MS Analysis Services ... The other two are more sequence analysis problems. ... > mining the apache log file for my final year project. ...
    (microsoft.public.sqlserver.dts)
  • Re: Dealing with Web Mining on Apache Log
    ... I've tried the trail version of the Coommerce Server, ... DTS to load web log ... MS Analysis Services ... The other two are more sequence analysis ...
    (microsoft.public.sqlserver.datamining)
Loading