Diversity Indices in Combining Classifiers for Cancer Diseases Data
Gebriel M. Shamia, Mufid Elmistiri, Mohamed A. Bosheha, Lecture · 2012
The behaviour of cancer diseases data from the Benghazi Cancer Registry as far as ecology through Broken-stick model for the spread of different species was used to study how diversity of the distribution of species of cancers is reflected in clusteringof zones and in different diversity measures related to such clustering. The aim of this paper is to study the dissimilarity in cancer distribution in the departmental divisions, among which are evenness component by , J Sp and species richness component by , R utilizing the generalized diversity index (GDI), , H , at the actual values , . The result from this registry data showed that an interrelating in diversity among divisions in the areas (zones) of rare and common of species is exist in five clusters. It is indicated that the most powerful tools for discriminating the spatial variations of cancer species diversity are in the multivariate category and its superimposition with cluster analysis was recommended in order to obtain more information regarding the relationship between sites (zones). However, ordination by diversity measures is a more informative summary than cluster analysis for such data.