An Approach to Building a Distributed ID3 Classifier

摘要

Current applications from industry, science, and business are storing hugeamount of data everyday. This data most of the time comes from distributed sourcesand are usually analysed for the organizations to discover knowledge and recognizepatterns by means of Data Mining (DM) techniques. This analysis usually requires toput all information together in a big centralized datasets. Analysing this huge datasetcould be very expensive in terms of time and memory consuming. For reducing this costsome Distributed Data Mining (DDM) architectures have been developed in recentlyyears. This paper presents an approach to building a distributed ID3 classifier whichtakes only metadata from distributed datasets avoiding the total access to the originaldata. This approach reduces the computing time nedeed to build the classifier.