Research On The Analysis Method Of The Personal Social Relationship For The Experts Avoidance Of The Science And Technology Appraisal

With the development of information technology, social software is ceaseless emerge in large numbers, such as Facebook and Twitter. More and more people used the software to making friends and interactive, many virtual friendships is emerged. Friends recommended and community detection to social network have become a research hotspots. But in addition to these virtual friends, the mining and discovery of the real social relationship in life has greater significance and value.Science and technology appraisal system as a network appraisal method is different from the traditional appraisal method of telephone and conference, the approval, selection and management of science and technology award projects are completed online through the system, it’s involved the avoidance process that the relationship of appraisal expert and the project applicant is close. This paper studies the application of social relations in the science and technology appraisal in view of the particularity of the experts avoidance in the appraisal system of science and technology award. O n the basis of the current extraction and application of personal social relationship, the main research contents include the following aspects:(1) Be aimed at the tagged corpus deficiency of the personal social relationship, a simple way to label the personal social relationship is used to build a corpus through the collection of Web pages, the word segmentation, part of speech tagging and other text preprocessing, a total of eight major categories of social relationships is labeled. By introducing the method of CHI feature selection in text categorization, only verbs and nouns in the text are selected as the feature items, which can effectively reduce the dimension of feature vectors. SVM is used for classification, experiments show that the feature selection method can improve the accuracy rate and recall rate of social relation extraction in a certain extent.(2)Most Webpages are related to people’s name or other information, but most of the information is dispersed and unstructured. According to the characteristics of Chinese expression, extracted the personal attributes of birthplace, graduate institutions, work units and achievements by construction of the trigger words based on the method of rules. Experiments show that the method based on the trigger word and semantic information is effective in the extraction of the personal attributes. In this paper, through the extraction of the personal attribute information, the potential relationship s between the persons are found.(3)The applications of personal social relationships are widely used, not only can through the personal social relationships for community discovery, advertising push and products recommended, also can used to mining the concerned personal information of the stars and other public persons show to the users. This paper puts forward a kind of person information model to organized the information effectively and divided the relationship between the characters into two categories of the main relationships and the secondary relationships. According to the characteristics of expert avoidance in the appraisal of science and technology award, an expert avoidance model based on the personal social relationship is proposed. The model can be used in accordance with the differences between personal social relationships and the information sources as far as possible to avoid the expert who knows or may know the project declaration. According to the analysis of the theories and methods, the model has strong feasibility.