Font Size: a A A

Text Keyword Extraction Analysis Platform Based On The Complex Network

Posted on:2018-11-02Degree:MasterType:Thesis
Country:ChinaCandidate:M X XuFull Text:PDF
GTID:2348330536979973Subject:Control engineering
Abstract/Summary:PDF Full Text Request
With the development of the Internet,keyword becomes an essential tool for searching information as it summarizes the text topic.Nowadays,how to mining keyword quickly and efficiently becomes a research focus.And scholars are keen on the way based on complex network as it is the latest keyword extraction methord.This paper transfers text into a complex network to do further research and analysis.In the meantime,a keyword extraction and analysis platform is constructed to realize the automatic keyword extraction.The main contributions have been summarized as follows:1.Reviewed the research situation of domestic and foreign scholars on the keyword extraction,this paper introduces the classic methods of keyword extraction in different areas and analyses the limitation of them.Aiming at the existing algorithm of keyword extractiom based on complex network,this paper introduces and analyses the measurement of important nodes commonly used in complex network,including frequently used statistical parameters and algorithms.2.Considering the importance of term frequency to the text theme,the concept of “term-frequency-shared weight” is put forward,and then a new method of constructing weighted text network is proposed.Through allocating the term frequency value of target node to its edges according to the importance of neighbor nodes,this method constructs a new weighted network which is different from the one built in current way,which is based on “the number of words' co-occurrence frequency in the same sentence”.3.On the basis of weighted text networks using term frequency character,the location weight coefficient is introduced considering of the character of human language.Then a keyword extraction method based on complex network named LTWPR is put forward.By the multi-class keyword extraction experiments of Sina news corpus and the comparison of experimental results with two classical algorithms,LTWPR algorithm is proved to be accurate and effective.This paper shows that LTWPR algorithm is excellent in mining keyword and is suitable for key nodes mining of bulk text networks.4.In this paper,a keyword extraction and analysis platform has been developed and showed,which achieves reading in texts and output corresponding keyword in bulk.The platform has the advantages of simple and friendly interface,convenient operation and strong expansibility.It can be used to simulate many kinds of keyword extraction algorithms and compare the results with the keywords marked by author.The keyword extraction and analysis platform integrates the research results of this paper,which is helpful to study the text keyword extraction quickly and intuitively.
Keywords/Search Tags:keyword extraction, weighted text network, location weight, complex network
PDF Full Text Request
Related items