Real Time Enhanced Random Sampling of Online Social Networks

TL;DR

This paper introduces a new real-time enhanced random sampling method for large online social network graphs, addressing access limitations and resource constraints, and evaluates its effectiveness on Twitter data.

Contribution

It proposes a novel sampling approach tailored for large, restricted social graphs, with comprehensive evaluation across multiple settings and graphs.

Findings

The method improves sampling accuracy under resource constraints

Different configurations show varying effectiveness depending on the graph

Best application scenarios are identified for the proposed sampling approach

Abstract

Social graphs can be easily extracted from Online Social Networks. However these networks are getting larger from day to day. Sampling methods used to evaluate graph information cannot accurately extract graph properties. Furthermore Social Networks are limiting the access to their data, making the crawling process even harder. A novel approach on Random Sampling is proposed, considering both limitation and resources. We evaluate this proposal with 4 different settings on 5 different Test Graphs, crawled directly from Twitter. Through comparing the results we observe the pros and cons of its method as well as their resource allocation. Concluding we present their best area of application.