Sign Prediction on Unlabeled Social Networks Using Branch and Bound Optimized Transfer Learning

Abstract

Sign prediction problem aims to predict the signs of links for signed networks. Currently it has been widely used in a variety of applications. Due to the insufficiency of labeled data, transfer learning has been adopted to leverage the auxiliary data to improve the prediction of signs in target domain. Existing works suffer from two limitations. First, they cannot work if there is no target label available. Second, their generalization performance is not guaranteed due to that fact that the solution of their objective functions is not global optimal solution. To solve these problems, we propose a novel sign prediction on unlabeled social networks using branch and bound optimized transfer learning (SP_BBTL) sign prediction model. The main idea of SP_BBTL is to use target feature vectors to reconstruct source domain feature vectors based on relationship projection, which is a complicated optimal problem and is solved by proposed optimization based on branch and bound that can obtain global optimal solution. With this design, the target domain label information is not required for classifier. Finally, the experimental results on the large scale social signed networks validate the superiority of the proposed model.

Link-Sign Prediction in Signed Directed Networks from No Link Perspective

“…As pretrained models are more and more popular in natural language processing community and achieve a lot of success, transfer learning has been used in sign-prediction as well. This direction has a lot of potential usage as it does not require labelled data [66].…”

W-Index: An Index for Evaluating Link Prediction considering Only the Role of Wins

“…Link prediction is one of the most fundamental problems of complex networks, which aims to infer the network link formation process by predicting missed or future relationships based on currently observed links [1]. Being able to effectively and efficiently predict unobserved links will allow us to mine future interactions among members in the social network [2,3], conduct successful recommendation in useritem bipartite networks [4,5], provide guidance for the planning process of infrastructure systems [6], optimize asset allocation in the stock market [7], discover drug-target interactions or identify targeting drugs for new targetcandidate proteins [8,9], and save costs and reduce operational risks for manufacturing companies [10,11]. Up to the present, various methods have been proposed for link prediction [12][13][14][15][16], most of which can be classified with heuristic-based approaches and learning-based approaches [17].…”