Analysis on web public opinion orientation based on extending sentiment lexicon[J
- Yang Chao
- Journal of Chinese Computer
The rapid development of network had led to network information very mixed and some information "hidden" on the network. In order to create a healthy network environment, efficient discovery of sensitive information has become a priority. In this paper, we use promiscuous mode to receive data packets and propose a multi-thread and super pipeline mode reorganization method for data packets. We also present a method for web sensitivity discrimination based on Semantic Sentiment Classification. This method not only had been improved by combining with the features of web page, but also expanded the NTUSD resources and built a modified dictionary. On this basis, we introduce the match-window technology to determine the sensitivity more accurately. The experience shows us that this method is accurate and valid.