A Survey of Neighbourhood Construction Models for Categorizing Data Points

Abstract

Finding neighbourhood structures is very useful in extracting valuable relationships among data samples. This paper presents a survey of recent neighbourhood construction algorithms for pattern clustering and classifying data points. Extracting neighbourhoods and connections among the points is extremely useful for clustering and classifying the data. Many applications such as detecting social network communities, bundling related edges, and solving location and routing problems all indicate the usefulness of this problem. Finding data point neighbourhood in data mining and pattern recognition should generally improve knowledge extraction from databases. Several algorithms of data point neighbourhood construction have been proposed to analyse the data in this sense. They will be described and discussed from different aspects in this paper. Finally, the future challenges concerning the title of the present paper will be outlined.