Survey on Semantic-Based Organization and Search Technologies for Network Big Data
Chun Wu · Chinese Journal of Computers · 2015
With the development of information technology,massive data resources with heterogeneous structure appear in the cyberspace,which is known as the network big data and has attracted extensive attentions.For mining the useful information from the network big data,it is required to efficiently organize the data resources in the cyberspace and realize the semantic-based similarity search.For an efficient data organization and search,we firstly need to extract the features/attributes of the big data to construct its high-dimensional semantic space,then define the data resources and queries as feature vectors or high-dimensional points in the semantic space,and finally can calculate the semantic similarity by the distance of high-dimensional points or the cosines of the angles between feature vectors.The multidimensional indexes can efficiently organize data resources in the semantic space,realizing the semantic-based similarity search.In addition,the dimensionality reduction technology can avoid the effects ofcurse of dimensionalitywhen the dimensionality of the semantic space is too high.In this paper,the existing multidimensional indexes and dimensionality reduction technologies are reviewed systematically.Moreover,the existing semantic-based similarity search technologies using distributed technology are analyzed,and some suggestions about future research work are discussed.