Thesis work is tackling two different but strictly related aspects: i) the implementation and test of a web application for distributed data mining on massive data sets and ii) the application of the above to a series of Data Mining (DM) problems connected with the study of photometric and astrometric transients in large multiband astronomical surveys i.e. to what is currently labeled as “time domain” astronomy. Time domain astronomy is still a largely unexplored field and it is expected to become one of the most relevant sources of discoveries in the next decade. New telescopes such as, for instance, the LSST will produce up to 10⁷ events/night and the detection, analysis and understanding of these events is posing tantalizing problems (both scientific and technological) to the astronomical and computer sciences communities.
For what DM is concerned, my work takes place in the framework of the DAME (Data Mining and Exploration) Project run jointly by the University Federico II, the California Institute of Technology and INAF Capodimonte Astronomical Observatory. The main goal is the implementation of an innovative platform (accessible as web application) for data mining in a distributed (S.Co.P.E.) computing environment of Massive Data Sets (multi-TB). In this context I have the responsibility of the implementation of the DM plugins and of the overall debugging. At the moment the package is in its alpha-version and can be accessed at the URL: http://voneural.na.infn.it/alpha_info.html.