A spectroscopy of texts for effective clustering

For many clustering algorithms, such as k-means, EM, and CLOPE, there is usually a requirement to set some parameters. Often, these parameters directly or indirectly control the number of clusters to return. In the presence of different data characteristics and analysis contexts, it is often difficult for the user to estimate the number of clusters in the data set. This is especially true in text collections such as Web documents, images or biological data. The fundamental question this paper addresses is: ldquoHow can we effectively estimate the natural number of clusters in a given text collection?rdquo. We propose to use spectral analysis, which analyzes the eigenvalues (not eigenvectors) of the collection, as the solution to the above. We first present the relationship between a text collection and its underlying spectra. We then show how the answer to this question enhances the clustering process. Finally, we conclude with empirical results and related work.

Silly us! Of course the grid doesn`t work : reading methodologies and policy texts on principals' work

This symposium presents work in progress from an ARC (discovery grant) funded investigation of principal supply, conducted by Jill Blackman, Judyth Sachs and Pat Thomas. Our research goals are to examine claims of an impending shortage of school principals in particular schools and localities, critically evaluate a range of possible reasons for this shortage, and ultimately, through woprk with principals' organisations, to develop some possibilities for policy action. In this symposium we focus on: (1) existing studies of principal supply (2) trends apparent from demographic and employment data, and (3) a text and interview based study of 'human resources' policy. We invite discussion on the implications of this first stage for the next - a national survey and interviews with teachers in pre-service training and in their first years of teaching.