Date of Award

10-30-2020

Publication Type

Master Thesis

Degree Name

M.Sc.

Department

Computer Science

First Advisor

Ziad Kobti

Keywords

Link prediction, Sign prediction, Signed social networks

Rights

info:eu-repo/semantics/openAccess

Creative Commons License

Creative Commons Attribution-Noncommercial-No Derivative Works 4.0 License
This work is licensed under a Creative Commons Attribution-Noncommercial-No Derivative Works 4.0 License.

Abstract

People hold all kinds of positive and negative feelings for one another. Social networking online serves as a platform for showcasing such relationships, whether friendly or unfriendly, like or dislike, trust or distrust, cooperation or dissension. These types of interactions result in the creation of signed social networks (SSNs). The sentiments among social individuals are complexity and diversity, and the relationships between them include being friendly and hostile. The positive (“friendly”, “like” or “trust”) or negative (“hostile”, “dislike” or “distrust”) sentiments in the relations can be modeled as signed connections or links. The missing relations or sentiments between individuals are always worthy of speculation. Hence, we need to predict negative sign prediction. Although negative signs typically dominate the positive signs in various analytical decisions in most real applications, it cannot be directly propagated between users like positive signs. The study on negative sign prediction is still in its early stages. There is a difference between the value of negative signs and the availability of these links in real data sets. It is therefore normal to analyze whether one can automatically predict negative signs from the widely available social network data. In this thesis, we propose a novel negative sign prediction model which includes negative sign related features from various categories to predict negative sign in signed social network. An extensive set of experiments is carried out on real-world social network datasets which demonstrate that the proposed model outperforms the existing method in predicting negative signs in terms of accuracy and F1 score(is a measure of a test’s accuracy) by 3% ∼ 4% and 5% ∼ 15% respectively.

Share

COinS