A Distributed Calculation Scheme for Contents Categorization

This paper describes a distributed calculation scheme for scoring relationship among documents. This scheme categorizes documents by using an algorithm which calculates a score value for the relationship between a category and a word in a document. The longer calculation time becomes when increasing the number of documents. Therefore, our scheme uses multiple machines. A master node divides a document set into several subsets, and it distributes them to each calculation nodes. Using this distributed calculation makes the calculation time short, and also makes the memory usage low.