12 September 2005 meeting mainly focused on giving feedback on project and planning the next action. from the last meeting till now i had to focus on getting the algorithms working, build models with distinct settings then plan on how to apply the models to new data and deciding on what kind of results would be expected. All this has been done. despite the updating or feedback, the meeting had more emphasis on apply the models to knew dataset, specifically the datasets obtained from the Centre for AIDS Development, Research and Evaluation Institute for Social and Economic Research. Basically, this brings a situation were we could kill two birds with one stone. We could provide answers to the Centre for AIDS Development, Research and Evaluation Institute for Social and Economic Research to their research area with the results we obtained from their data at the same time with me evaluating the algorithms. plan of action. 1) apply the models with the Centre for AIDS Development, Research and Evaluation Institute for Social and Economic Research in mind, ie, considering their requirements. 2) also have a critical look at the the other algorithms available in oracle data miner 10g namely, Predictive variance (attribute importance) and Apriori (association rules). this would however shift the researchs' path to something else.
----------------------------------------------------------------------------------------------------------------------------------- 29 August 2005 currently i am having some problems with the K-means algorithm building a model with the dataset that i obtain from the Centre for AIDS Development, Research and Evaluation Institute for Social and Economic Research which is obviously AIDS related. the problem was that the algorithm gives errors during the build, and these errors were due to memory shortage on the system. this meeting was jus a short meeting do discuss about this and try and look for possible ways to go round this problem. After close examination of the problem, John and i decided the increasing the systems physical memory would be an option, but currently the system is running on 1gig of RAM which is quite large, so increasing it would probably make the system unstable. therefore what we finally decided on was to partition the dataset into smaller sections: so plan of action was--> partition dataset into smaller and reasonable size then build models and see if both clustering algorithms are working fine. ---------------------------------------------------------------------------------------------------------------------------------- 15 August 2005 this meeting was mainly at getting feedback from John on the draft of my project poster which was due on the 22nd of the same month for submission to e printed. Also included was feedback on project progress, that is, how the instalations are going. on that note, everything is looking good, software working quite well. ----------------------------------------------------------------------------------------------------------------------------------- 11 August 2005 During this meeting, i was seeking some advice from John on what to do as i was having problems connecting the data mining software to the oracle 10g database during the installation of the oracle Data miner software. at this meeting John suggested that i contact the Oracle technicians via oracle-metalink which is n Oracle's online support service. Plan B, was that if i completely failed to make the data mining with Oracle10g i work on the older version oraclei. ----------------------------------------------------------------------------------------------------------------------------------- 4th August 2005 Although the data mining software was not installed at the present time. i found it critical to at least have the datasets for the datamining available. On this day, John and i went out for data hunting. we approached the Centre for AIDS Development, Research and Evaluation Institute for Social and Economic Research, and met Kevin Kelly whom we had to convince that we needed there datasets and maybe in that note could possibly find patterns that they could find useful. This was possible with data mining since it is strongly believed that data mining could retrieve information that statistical tecniques cannot. ----------------------------------------------------------------------------------------------------------------------------------- 20 July 2005 this meeting was all about giving feedback on the developments during the vacation in terms of project developments, database installations and so on. As an oral presentation was due on the 25th of the same month, John gave me some guide lines on how to deliver the presentation as this was critical since the majority of the computer science staff will be present, they needed to know the progress of my project. ----------------------------------------------------------------------------------------------------------------------------------- JUNE/JULY VACATION: during this time, i managed to install Oracle10g database on the server Athena. Initially, the plan was to install it on Ora1 but due to the two databases conficting ie oracle9i and 0racle10g lead to the installation on a different server. 17 May 2005 This meeting was all focused on the Literature review which was meant to be submitted on the 23rd of the same month. The previous day before the meeting, i submitted a draft of my literature review. During this meeting thats were i got a feedback on what i should include or remove. ----------------------------------------------------------------------------------------------------------------------------------- 8 May 2005 Discussions on the Oracle10g Licence, downloading and installation. In terms of downloading, John made arrangements with the CS Tech guys to have the database software dowloaded as the whole software was about 2gigabytes. In terms of the installations, I agreed with John to have the database software installed sometime during the June vacation on the server machine Ora1 which also contains the database oracle9i. ----------------------------------------------------------------------------------------------------------------------------------- 2 May 2005 This meeting was mainly to advise John on any developments, difficulties ,that is, problems and and suitable sudjestions. At this meeting, gave me a few guideline on how to go about writing the literature review. ----------------------------------------------------------------------------------------------------------------------------------- 8 April 2005 Had a number of short unscheduled meetings with John mainly to keep me on track as to making sure that i understand what i am doing. Advised me to do as may data mining tutorials as possible in order to get a better and more practical understanding of the field. On this scheduled meeting, John gave me some guide-line on my first presentation which was mainly focused on the project poposal ----------------------------------------------------------------------------------------------------------------------------------- 10 March 2005 I met with John Ebden and submitted my draft of the project proposal, then we met again after two hours for feedback. At same meeting he handed me the server that I would be working with [Ora1]. This machine is installed with the oracle9i database which embeds the data mining software ODM 9i. Unfortunately, for my project i dicided to work with oracle10g which is the latest database release from oracle. We also agree to move the machine next to my machine in the honours lab because he feared that when I want to access the machine it may be off. And in cases when he is not available this would be a great inconvenience on my part.
----------------------------------------------------------------------------------------------------------------------------------- First Meeting with Supervisor Meeting with Supervisor-1 March 2005. Met John Ebden on the 1 st of March 2005 and he gave me more detail about what the project is all about and what I am expected to do. This was kind of a introductory meeting were he gave me a number of textbooks to read to have a general background on the field.
|
|---|