Font Size: a A A

Research On Forecasting Models And Platform Of Online Public Opinion

Posted on:2010-03-04Degree:MasterType:Thesis
Country:ChinaCandidate:J ZhangFull Text:PDF
GTID:2189360275973405Subject:Communication and Information System
Abstract/Summary:PDF Full Text Request
The Fourth Plenary Session of the Sixteenth Central Committee of the Party pointed out that we should establish and improve social public opinions' collection and analysis mechanism and keep the avenues open for getting to know the social conditions and public opinions, which manifests that the Central Party Committee attaches great importance to the research of public opinion. In recent year, the researches on public opinion in the domestic circle are getting more and more profound and have come out with many great results. However, there are still many fields need to be studied. Along with the network technique applied to the universality and its development, fundamental changes have taken place in the spreads way and speed of the public opinion. Online public opinion influenced the human social state, boosts the surveillance about the official and decision-making, and has the following characters: spreading gossip too widely to control effectively, spreading information diffusely, rapidly with grave influence, and the unreasonable and emotional tendency of the information communication.In order to strengthen the supervision and direction of online public opinion, we should pay attention to the collection and analysis of the information to maintain the social stabilization and establish early warning mechanism of online public opinion. We took the research object of actual network forum, Blog channel and news website in this paper, and proposed a forecasting platform construction scheme, which provided a significant exploring method, based on Java and various forecasting models.Platform using the optimization forecasting model to forecast the original time series formed from reply numbers of forum post, personal log and news report can timely, exactly and conveniently provide hot topic development trends to internet manager. The universal historical data collection and preprocessing were stated from this paper completely, which plays an important role in realizing platform. This paper includes below major innovative point.The forecasting model is the very important aspect of building such platform. At present, we don't have suitable forecasting tools. To deal with this phenomenon, a fire-new thinking is introduced. The calculating process of each forecasting model about economic research field is analyzed at detail and they are used to do prediction for online public opinion. Through a comparative analysis, the estimation results from these forecasting models can be acceptable.
Keywords/Search Tags:online public opinion, short-term forecasting, single exponential smoothness, ARIMA model, BP neural network
PDF Full Text Request
Related items