Friday, August 3, 2007

Thesis Update

This week I have outlined a few approaches to use in analysing the data once a data set has been decided upon. Firstly, I will use Fuzzy Clustering, and then Switching Regression, as outlined by Dr Poon, in order to see how I.T. correlates to Industry Structures. A second method I will be attempting, if time permits, would be to perform Switching Regression first, and then perform Fuzzy Clustering, in the method outlined by Hathaway and Bezdek in their 1993 paper titled Switching Regression and Fuzzy Clustering published in IEEE Transactions on Fuzzy Systems, Vol 1, No 3, August 1993,

I will attempt these two methods using firstly the "cor" and "fanny" functions for R which I have in literature on the internet (I have not tested these two functions as of yet, but will do so this week). However, this may only provide limited support for regression and clustering, and I may have to write the scripts to perform these functions in R, as I wish to do both Fuzzy C-Means, and Fuzzy C-Regression for the Fuzzy Clustering part of this project.

Additionally, we (Dr Poon's thesis group), have attempted to compile a data set in the past week, which we will be refining and adding to in the coming week.

No comments: