Swarms on continuous data
Vitorino Ramos, Ajith Abraham · 2004
While being it extremely important, many exploratory data analysis (EDA [J. Tukey (1977)]) systems have the inability to perform classification and visualization in a continuous basis or to self-organize new data-items into the older ones (even more into new labels if necessary), which can be crucial in KDD - knowledge discovery [U.M. Fayyad et al., (1996), (1992)], retrieval and data mining systems [S. Mitra et al., (2002), U.M. Fayyad et al., (1996)] (interactive and online forms of Web Applications are just one example). This disadvantage is also present in more recent approaches using self-organizing maps [R. Brits et al., (2001), H.P. Siemon et al., (1990)]. On the present work, and exploiting past successes in recently proposed stigmergic ant systems [V. Ramos et al., (2002)] a robust online classifier is presented, which produces class decisions on a continuous stream data, allowing for continuous mappings. Results show that increasingly better results are achieved, as demonstrated by other authors in different areas [V. Ramos et al., (1999), A. Lumini et al., (1997)].