Research On Micro Blogging Based On The Android System

Microblogging is very important in circles, in circles, we are each other’sattention, the topic generated the correlation is relatively high. Circle of theconstruction technology and personalized recommendation are closely related.Personalized recommendation friendly determines the direct experience of the userson the microblogging. Personalized recommendation using the most widely usedmethod is to cluster analysis. However, the traditional long text clustering algorithmfails to consider the microblogging data sparsity characteristics of the ordinaryshort-text clustering algorithm has failed to combine the characteristics of themicroblogging data, such as topic-related.Microblogging feature of the data, this paper proposes a microblogging vectorconstruction method, consider the topic of microblogging and microblogging textcontent, calculating their similarity based on vector space model, respectively, fromthe text of the topic and microblogging. Microblogging text similarity computing thesimilarity formula has been adjusted, when the microblogging contains words relatedto the topic, the words dramatically. Finally, the normalized topic similarity andmicroblogging text similarity, similarity of the two microblogging.The short version of the microblogging feature of the data, this paper presents amicroblogging clustering algorithm. Discovery algorithm on the basis of similaritycalculation, according to the vocabulary of the proposed label, labeling vocabularycontinuous microblogging data set is divided into different clusters, making a clustercontains the tag vocabulary, another cluster does not contain a tag vocabulary.Repeating to use of the iterative segmentation method, the final gets a good cluster. Inthis paper, part of the data in the corpus of NLPIR microblogging content tested theMBCA algorithm, evaluation criteria FScore value, with open source clustering toolpackage CLUTO in other clustering algorithms were compared.On the basis of the above work, it has designed and relizationed of a simplemicroblogging system based on the Android platform using the the MBCA clusteringalgorithm. Software running on a real machine simulator based on the variousversions of the Android platform and ME525in good condition, the recommended listof appropriate proof of good software design, and also validating the rationality of theMBCA algorithm.