Font Size: a A A

Research On Resource Semantic Space And Retrieval Of Scientific Literature

Posted on:2020-03-30Degree:MasterType:Thesis
Country:ChinaCandidate:Z J LiFull Text:PDF
GTID:2428330602952148Subject:Information Science
Abstract/Summary:PDF Full Text Request
As an important academic information resource,scientific literature is the preferred reference resource for researchers in the research process.With the advent of the era of big data,the network is full of huge amounts of scientific literature,and maintains a rapid growth momentum.Therefore,how to achieve efficient retrieval of internal knowledge of scientific literature has become the focus of the research.At present,the retrieval system of scientific literature mostly adopts the method of coarse-grained knowledge organization based on the whole document,which fails to reveal the fine-grained knowledge content and its relationship within the scientific literature.In the retrieval process,the retrieval system can only receive the keywords combination form of retrieval,and can only match the retrieval form with the specific field of the scientific literature,which means that the retrieval process lacks semantic analysis,resulting in the retrieval results can not meet the real needs of users.Therefore,how to realize fine-grained organization and semantic retrieval of internal knowledge of scientific literature is the research focus in the field of scientific literature retrieval.In view of this,from the perspective of fine-grained knowledge unit,this paper attempts to study the fine-grained knowledge organization method and semantic retrieval method of scientific literature.The main work of this paper includes two aspects:On the one hand,a fine-grained knowledge organization method for scientific literature is proposed,namely,Resource Semantic space.Guided by the idea of fine-grained knowledge organization and taking scientific literature as the research object,this method extracts the knowledge units with the ability of complete knowledge expression by analyzing the internal logical structure of scientific literature.On this basis,the semantic link network between knowledge units and between knowledge units and their internal concepts is established by using the technology of semantic link network construction.Finally,a fine-grained knowledge organization method for scientific literature is proposed,namely,the resource semantic space,which includes the resource layer,the description layer and the semantic link layer of knowledge units.This method can lay a foundation for fine-grained semantic retrieval research.On the other hand,a fine-grained semantic retrieval method for scientific literature based on resource semantic space is proposed.On the basis of fine-grained organization of scientific literature,this study proposes a method for calculating the semantic similarity between retrieval and knowledge unit theme concepts by analyzing the semantic dependence,and a method for calculating the semantic similarity between retrieval and knowledge unit element concepts by analyzing the co-occurrence of words.And finally,a matching method between natural language retrieval and knowledge unit is proposed.The experimental results show that the method can directly locate the knowledge content within the scientific literature,and has a high accuracy.From the perspective of fine-grained knowledge unit,a new method of organization and retrieval of scientific literature is proposed.At the theoretical level,this method refines the granularity of knowledge organization and retrieval objects,enriches the theory and method of fine-grained knowledge organization and retrieval;at the practical level,this method is conducive to revealing the internal and external semantic relationship of scientific literature,realizing fine-grained organization of internal knowledge units of documents,and improving the efficiency in accessing and utilizing information.
Keywords/Search Tags:scientific literature, fine-grained knowledge organization, knowledge unit, resources semantic space, semantic retrieval
PDF Full Text Request
Related items