Aiming at the problem of automatic proofreading of sensitive information in mass text content, an automatic proofreading method based on the combination of rule and SVM (Support Vector Machine) is proposed. To classify sensitive information based on important sensitive information provided in the "Newly Prohibited Texts and Cautions in Xinhua News Reports"(newest revision) and related central and online texts.According to the different categories, the paper constructs the classification processing rule base, designs the corresponding rules automatic Processing algorithm, and realizes the sensitive information automatic proofreading, At the same time, using the SVM model to analyze the result of the rule processing with emotion, which greatly reduces the false alarm rate. The test result shows that the recall rate of method is 89.98%, the accuracy rate is 98.31%, and 100,000 + text content is processed per second, which solves the key difficult problems in the practical engineering application.
展开▼
