Measuring Social Relations in the Web based on Search Engine and Text Analysis: a Model and Implementation

Abstract

To improve the accuracy and stability of existing methods to measure social relations in the Web, a novel model of relation measuring methods based on both a search engine and text analysis is proposed. The model measures the strength of social relations according to both the co-occurrence of two persons’ names in web pages obtained by a web search engine and the cooccurrence of two person names in sentences of the text of web pages as found by text analysis. The formalized description of the model is then presented. To evaluate the effectiveness of the proposed model, the specific implementation of the model is presented in detail. Experimental results show that compared with existing methods only based on a search engine or text analysis, the relation weights obtained by the specific methods based on the proposed model are more accurate and stable.