Modelling on classification and retrieval strategy in map-reduce based IR system

Abstract

Classification is essential to many web applications. Instead of the traditional segmentation as usual, this paper presents a novel non-segment classification approach, which can meet the requirement for big data classification & application. The novel algorithm removes the complicated time consuming segmentation. The experimental results and the analysis show the feasible of the approach. Based on the proposed approach, a map-reduce based distributed IR system is present.