A fuzzy methodology for clustering text documents with uncertain spatial references

Abstract

Fuzzy ERC (Extraction, Resolving and Clustering) architecture is proposed for handling the uncertain information that can be either queried explicitly by the user and the system can also cluster the documents based on the spatial keyword present in them. This research work applies fuzzy logic techniques along with information retrieval methods in resolving the spatial uncertainty in text and also finds the spatial similarity between two documents, in other words, the degree to which two or more documents talk about the same spatial location. An experimental analysis is performed with Reuter’s Data set. The results obtained from the experiment are based on the empirical evidence of the document clustering based on the spatial references present in them. It is concluded that the proposed work will provide users a new way in retrieving documents that have similar spatial references in them.