Research On Predication In Popularity Of The Public Sentiment Of News Information Based On Neural Network

Data forecast refers to the analysis of data to estimate or speculate about the future of the process on the basis of existing data. With the rapid development of big data and cloud computing, data processing, analysis and prediction methods has drawn increasing academic attention. With the rapid development of smart phones and the widespread use of social networks, various network news is growing at an exponential growth rate of the rapid spread on the Internet. News Network with its own easy to produce, easy to copy, short length, short life cycle and low cost characteristics, to become the first mobile phone users browse resources and major social networking sites. We found that users welcome news in the mass of news, and push its orientation has become a new hotspot circles.Internet news public opinion research popularity (popularity degree) has important practical significance. Article mainly from the popularity of the two aspects of the news public opinion:1. For specific sampling and online news properties feature extraction and feature vectors extracted correlation analysis, based on principal component analysis and factor analysis. Using principal component analysis to select from a number of important influence factors unrelated features, many of the original property dimensionality reduction.2. Discussion of the prediction algorithm and prediction BP neural network based on public opinion in the news area. Considering the popularity of high-dimensional prediction model using BP neural network on the main feature vectors for training, the establishment of a news popularity of predictive models. The simulation results show, BP neural network model can accurately predict the popularity of news, it is possible to achieve higher effective in different real environmental prediction accuracy.