Abstract
In this paper we study a new technique we call post-bagging, which consists in resampling parts of a classification model rather then the data. We do this with a particular kind of model: large sets of classification association rules, and in combination with ordinary best rule and weighted voting approaches. We empirically evaluate the effects of the technique in terms of classification accuracy. We also discuss the predictive power of different metrics used for association rule mining, such as confidence, lift, conviction and chi(2). We conclude that, for the described experimental conditions, post-bagging improves classification results and that the best metric is conviction.

Abstract
The basic principles of a methodology for remote collaborative data mining are proposed. Starting from CRISP-DM, a general data mining process designed to carry out data mining projects; it is described how the principles of knowledge sharing and ease of communication can be embedded in the data mining process, The aim is to allow the execution of data mining projects, with the participation of multiple experts working from distant locations. All the participants in such a project can profit from the knowledge produced by others and share their knowledge online with the other participants. The produced knowledge (for example data transformations, working hypothesis, models, results of experiments) is also stored for future inspection and use, in pursuit of organizational learning. A prototypical implementation (RAMSYS) of the remote collaborative methodology is described with examples.

Abstract

Abstract

Abstract

Abstract