AN INNOVATIVE APPROACH TO TIME COMPLEX DATA STRUCTURE FORMATION VIA REVERSE NEAREST NEIGHBOUR DENSITY CLUSTERNG
Keywords:
clustering, analysis of similar patterns, time-sensitive searches, density-based clustering, influence border clustering, and nearest neighbor search..Abstract
Based on prior knowledge of data sets pertaining to different aspects, data mining research has introduced various clustering, unsupervised, and machine learning-related methodologies, algorithms, and approaches to extract similar objects from diverse data sets pertaining to different categories. Many search engines, including those that scour news archives and blog posts, use time as a relevant dimension. Finding documents that are topically similar to a query has been the primary focus of research on searching over such collections so far. Ranking documents based on subject similarity alone has its limitations, however. We find that, in addition to subject similarity, the publication time of documents in a news archive is essential for a significant class of queries we refer to as time
sensitive enquiries, and that these factors should be considered when determining the final document ranking. Improving retrieval for "recency" searches, which target recently created documents, has been the primary focus of previous work.
References
Yinghua Lv, TinghuaiMa, "An efficient and scalable density-based clustering algorithm for datasets with complex structures", Neurocomputing171(2016)9–22.http://dx.doi.org/10.1016/j.neucom.2015.05.109.
H. Zhou,X.Wang,X.Zhao, An efficient density-based clustering algorithm combined withrepresentativeset,J.Inf.Comput.Sci.10(2013)2021–2028.
Y.H.Lu,T.H.Ma,S.M.Zhong,J.Cao,X.Wang,A.D.Abdullah, Improved locality sensitive hashing method for the approximate nearest neighbor problem, Chin. Phys.B23(2014)080203.
