Expert evidence document recognition integrating relationship of expert

摘要

Focus on the issue that the existing expert evidence document recognition method is not effective use of the relationship between experts, this paper proposed a method to recognize the expert evidence documents which integrate the feature of webpage content and the complex relationship between the experts. First of all, the method analyses characteristics of expert evidence document, to select Coauthors and organization as relationship feature, and then extracts feature values. Secondly, select 10130 expert evidence pages, and build data sets which integrate the feature values of and webpage content and the relationship, and then select different learning algorithms to build recognition model. Finally, experiment of expert evidence document recognition is carried out in the field of information processing. The experimental results shows that the proposed method achieves a good recognition effect. Using relationship can effectively improve the accuracy of the evidence document recognition. Its accuracy is improved by 4% compared to method using the page content feature recognition only.