Nowadays, large volumes of commercials reach the desks of executives magnifying the potentials of this new discipline. This is probably the first time in which we recognize our need for a data mining tool, our questions being: "what have all these tools and services in common, which of them are market leaders and which of them offer the best functionalities?" Text data bases, for instance, like commercials, news, and summaries can be classified by programs that without a real understanding of the text can separate them into homogeneous groups. This is achieved through a combination of conceptual clustering and frequency analysis on words. This application is by no means exclusively related to telecommunications, everybody can rely on data providers for almost any kind of phenomena, from retail sites to goods movements among harbours. External data are a considerable mass and this is common to other industries as well, but probably none with the same amount as Telecommunications. Internal data are generated by all the transactions happening over the network. At this stage the only problem left seems to be the exploitation of all the available data.

The paper introduces ADHOC, a tool that integrates statistical methods and machine learning techniques to perform effective feature selection. Feature selection plays a central role in the data analysis process since redundant and irrelevant features often degrade the performance of induction algorithms, both in speed and predictive accuracy. ADHOC combines the advantages of both filter and feedback approaches to feature selection to enhance the understanding of the given data and increase the efficiency of the feature selection process. We report results of extensive experiments on real world data which demonstrate the effectiveness of ADHOC as data reduction technique as well as feature selection method. ADHOC has been employed in the analysis of several corporate databases. In particular, it is currently used to support the difficult task of early estimation of the cost of software projects.

JazzMatch, a parallel matching algorithm explicitly designed for secondary memory-based production systems, is presented. JazzMatch is a state-saving algorithm that performs incremental match. It exploits extremely fine-grained parallelism and optimizes the storage state by permitting the sharing of common conditions in the rules. JazzMatch is supported by a message passing parallel architecture. A cost and performance analysis of JazzMatch is provided, and the results are compared with those for other schemes in current literature.>