Danil Shaikhelislamov, Mikhail Drobyshevskiy, D. Turdakov, A. Yatskov, M. Varlamov, Denis Aivazov
{"title":"Three-step Algorithms for Detection of High Degree Nodes in Online Social Networks","authors":"Danil Shaikhelislamov, Mikhail Drobyshevskiy, D. Turdakov, A. Yatskov, M. Varlamov, Denis Aivazov","doi":"10.1109/IVMEM51402.2020.00014","DOIUrl":null,"url":null,"abstract":"This paper considers the problem of influential users detection in online social networks. Identifying of such key entities is of interest in many areas: marketing, politics, information security, business. The degree of the node of the corresponding graph is used as a popularity indicator in this work. Network query limitation is the main challenge in discovering their structure. Therefore, our task is to detect a percentage of the highest degree network nodes under a budget restriction. We propose a three-step crawling algorithm in two versions to solve the problem. We experimentally show its efficiency at various budget limits and superiority over known crawling strategies. For example, to detect top-1% of hubs with 90% precision, one needs to crawl 5% of graph nodes in average with our 3-StepBatch algorithm. We also show that our algorithm performs well for different target set sizes, from 0.01% to 10% of the graph.","PeriodicalId":325794,"journal":{"name":"2020 Ivannikov Memorial Workshop (IVMEM)","volume":null,"pages":null},"PeriodicalIF":0.0000,"publicationDate":"2020-09-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"1","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"2020 Ivannikov Memorial Workshop (IVMEM)","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.1109/IVMEM51402.2020.00014","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"","JCRName":"","Score":null,"Total":0}
引用次数: 1
Abstract
This paper considers the problem of influential users detection in online social networks. Identifying of such key entities is of interest in many areas: marketing, politics, information security, business. The degree of the node of the corresponding graph is used as a popularity indicator in this work. Network query limitation is the main challenge in discovering their structure. Therefore, our task is to detect a percentage of the highest degree network nodes under a budget restriction. We propose a three-step crawling algorithm in two versions to solve the problem. We experimentally show its efficiency at various budget limits and superiority over known crawling strategies. For example, to detect top-1% of hubs with 90% precision, one needs to crawl 5% of graph nodes in average with our 3-StepBatch algorithm. We also show that our algorithm performs well for different target set sizes, from 0.01% to 10% of the graph.