| Summary: | Grid Data Mining tools must be able to cope with very large, high dimensional and, frequently heterogeneous data sets that are geographically distributed and stored in different types of repositories, produced from different devices and retrieved through different protocols. This paper presents an agent-based version of a Learning Classifier System. An experimental study was conducted in a computer network in order to determine the systems’ efficiency. The results showed that the model is suitable to be applied in inherently distributed problems and is scalable, i.e., when the latency communication times are not considerable, the system obtains an interesting speedup.
|
|---|