Analysis Of Microblogging Information Classification

Microblogging as a new interactive network platform, contains vast amounts of information. In these information, contains a huge potential business opportunities.Microblogging, as an emerging network interactive platform, contains mass information which posses huge potential business opportunities. As a social networking platform, Microblogging requires enterprises providing to their customers or clients timely services. The enterprises have the responsibility and obligation to provide timely technical assistance to the customers who complained or doubted the products on their own Microblogs. Likewise, the enterprises have the responsibility and obligation to provide timely information and guide services on sale products to the customers who showed purchase intentions on their Microblogs.For any enterprise, mining and digging information through billions of Microblogs is the key to obtain valuable first-hand information. Meanwhile, it is important and noteworthy for enterprises to make sentiment classification on Mircoblogs information that related to their products and to decide whether technical support services or products shopping guide is necessary.This thesis will take the Lenovo Group as a corporate background. The thesis will make classification to the related microbloggings on whether they are relative to Lenovo or its products, on the emotional tendencies that the microbloggings revealed, and on the necessity for the shopping guide services or technical support services. The author will design a framework of information classification system in accordance with the functional requirements of the microblog. On the basis of the function that the information classification system would achieve, the author will design a classifier and stress on the design of a feature extraction algorithm which will be applied in a comparison experiment with the simple one. In order to detect the function and classification result of microblog information classification system and the classifier, the author will test the classifier for further improvement. The last part of this thesis will analyze the test results of microblogging information classification system. This paper will analyze the results of comparison test of the classifier in the feature extraction algorithm and simple to use feature extraction algorithm, and will draw a conclusion on which algorithm is better in classifying Micoblog information. When making separate analysis on feature extraction algorithm, the author will focus on the characteristics of recall, precision and harmonic value and analyze the specific effect of the classification. After analyzing the classification results and based on the tests results, the thesis will analyze the shortcomings of current classification system and provide optimization proposal on microblogging information classification system. Finally,this thesis will make an outlook on the microblogging information classification system.