The Construction And Optimization Of The Associated Semantic Representation Model For The Ultra-short Microblog Text

Under the circumstance of big data,the ultra-short microblog text is difficult to be independently understood because of its special characteristics such as context shortness,semantic sparseness,information fragmentation and so on.To solve this problem,the associated semantic representation model of ultra-short microblog text(ASRM-UMT)is proposed in this thesis to help users understand it better.First of all,the semantic content of ultra-short text can be extended by the constructed dynamic time window.After that,the associated semantic of ultra-short microblog text can be presented by the directed and weighted associated semantic network graph.At last,the ASRM-UMT can be optimized by the application of clustering coefficient in the network.The main research contents are as follows:(1)The dynamic time window can be constructed to extend the semantic of ultra-short microblog text.Aiming at the features of content shortness and semantic sparseness of the ultra-short microblog text,the way of constructing dynamic time window is proposed in this thesis to expand the semantic for ultra-short microblog,and to make the semantic of the ultra-short microblog text richer in special context.(2)The associated semantic representation model of ultra-short microblog text can be constructed.Based on the special context in the dynamic time window,the ASRM-UMT can be represented by a directed and weighted associated semantic network graph,and the associated relations between the feature keywords can be extracted by mining algorithm.(3)The associated semantic representation model of ultra-short microblog text can be optimized.Based on the constructed ASRM-UMT in specific dynamic time window,through expands the dynamic time window,the associated semantic representation model can be optimized by the computation and comparison of the clustering coefficients at some different dynamic time windows.The experimental results show that the ASRM-UMT can express the semantic sparseness' s ultra-short microblog text accurately and effectively.In this thesis,the extension of dynamic time window is proposed to enrich the semantic of ultra-short microblog text,the form of graph structure is adopted to present the semantic of the ultra-short microblog text,and the clustering coefficient is applied to optimize the ASRM-UMT.The research of this thesis can effectively express the semantic of ultra-short microblog text in the special content and help users to understand the meaning of ultra-short microblog text quickly.Meanwhile,it also can provide a theoretical basis for the text analysis of microblog platform.