Research and analysis of method of ranking micro-blog search results based on binary logistic model

Abstract

Ranking results in micro-blog search as user's interests is challenging because of the special form of micro-blog search results. To attempt to solve the problem, in this paper, we summarize the characteristics of micro-blog search results, propose a method using a sort of decision model- binary logistic model, test the confidence level of the model and estimate the weight of the variable in the model collecting the real samples. The result shows the relation between user's decision and the factors from each individual micro-blog search result as well as the feasibility of ranking using the model. We also analyze the model. © 2013 Springer-Verlag.