Research And Application Of Negative News Judgment Algorithm

Negative news are of great value to the Bank and Venture Company, but present search engine doesn’t provide the function to classify the emotional tendencies of the news. Artificial screening of news is inefficient and the workload is to heavy, so specify negative news relate to the searching key automatically is of great realistic significance and practical value.Under the basis of vast analysis and research of the text content in the Web, an algorithm combines the dependency grammar and the simplified case grammar was proposed, which is used to the negative news judgment by specific semantic analysis. After applying the algorithm to the engineering project, the result shows that it is useful and practical. The main task of this paper as follow:1. A method focus on extract the key sentences of the Web content is pot forward. This method is used to filter the noise text based on weighted graph and statistics theory, and extract the key sentences as the data for sentiment analysis.2. Based on the’HowNet’sentiment dictionary, this paper propose a new algorithm to identify the negative news by analyzing the interdependence of words through dependency parsing, and then filling in the template framework based on case grammar with the result.3. Apply the two algorithms to the engineering project based on the ’FudanNLP’development kit. The engineering project named "Negative News Automatic Retrieval System", has the following main functions:1) fetch the news from the mainstream search engine automatically.2) extract the title and key sentences from the news page automatically.3) analyze the key sentences of the news based on the algorithm this paper proposed, and mark the news as positive or negative.4. Experiments and primary application shows that the method and algorithm this paper put forward has good performance both in accuracy and efficiency, and fit for practical use.