FRINK Linked Data Fragments server

Matches in SemOpenAlex for { <https://semopenalex.org/work/W3161026078> ?p ?o ?g. }

• W3161026078 abstract "This thesis work considers a competition between two strategic agents who try to maximally influence a social network by targeting a finite number of non-strategic/regular agents. We assume that regular agents adjust their opinion through a distributed averaging process, whereas the strategic agents present a fixed belief, towards which they try to shift the average opinion of the overall network by identifying the optimal targets to connect to. More specifically, the competition is set from the perspective of one of the strategic agents, as the optimization problem of selecting at most k regular agents to connect to in order to shift the network average opinion. Such a problem is known to be computationally hard, and effective heuristics are needed to reduce its complexity. The core of this thesis work consists in the technical results upon which these heuristics are built. In particular, we tackle the problem in two ways. On one hand, we provide an alternative proof (with respect to the literature) of the submodularity of the objective function, so as to reduce the problem to a greedy heuristic -- in this way, rather than solving one computationally challenging optimal targeting problem, it is possible to find a suboptimal solution by solving k separate single targeting problems. On the other hand, we exploit the underlying graph structure to solve special instances of the problem, upon which we build more refined heuristics. As a first instance, we exploit the electrical analogy to provide the analytical solution for the single targeting problem over line graphs. Then, the peculiar behavior of the objective function over such a graph is proved to be extendable to generic tree graphs, by considering the tree branches as pseudo-line graphs. From these findings, we define an algorithm which finds the optimal solution in a much faster way with respect to a brute-force approach. Upon this result, then, we build heuristics that work for single targeting over sparse graphs. Nevertheless, as a second instance, we provide the analytical solution for the optimal targeting problem over complete graphs. The result here is huge, since it tells us in a simple way whether it is convenient or not to block the opponent's influence by targeting the same nodes. We then extend the same argument, with some modifications, to generic graphs in building the heuristics, generically leading the more accurate results with respect to the greedy approach. Upon these results, we combine blocking heuristics and greedy ones to get more accurate results, while we additionally combine them with degree based heuristics to save computational time. Finally, we compare on different special graphs and real social networks all these heuristics with respect to zero cost heuristics, built based on degree and blocking strategy. Summarizing the theoretical results, we obtain a scheme that tells which algorithm is more accurate and less computationally expensive, based on the underlying graph." @default.
• W3161026078 created "2021-05-24" @default.
• W3161026078 creator A5004417430 @default.
• W3161026078 date "2020-11-30" @default.
• W3161026078 modified "2023-09-26" @default.
• W3161026078 title "Optimal Targeting in Social Networks" @default.
• W3161026078 hasPublicationYear "2020" @default.
• W3161026078 type Work @default.
• W3161026078 sameAs 3161026078 @default.
• W3161026078 citedByCount "0" @default.
• W3161026078 crossrefType "journal-article" @default.
• W3161026078 hasAuthorship W3161026078A5004417430 @default.
• W3161026078 hasConcept C126255220 @default.
• W3161026078 hasConcept C127705205 @default.
• W3161026078 hasConcept C137836250 @default.
• W3161026078 hasConcept C138885662 @default.
• W3161026078 hasConcept C154945302 @default.
• W3161026078 hasConcept C165696696 @default.
• W3161026078 hasConcept C173801870 @default.
• W3161026078 hasConcept C177264268 @default.
• W3161026078 hasConcept C199360897 @default.
• W3161026078 hasConcept C33923547 @default.
• W3161026078 hasConcept C38652104 @default.
• W3161026078 hasConcept C41008148 @default.
• W3161026078 hasConcept C41895202 @default.
• W3161026078 hasConcept C51823790 @default.
• W3161026078 hasConcept C521332185 @default.
• W3161026078 hasConceptScore W3161026078C126255220 @default.
• W3161026078 hasConceptScore W3161026078C127705205 @default.
• W3161026078 hasConceptScore W3161026078C137836250 @default.
• W3161026078 hasConceptScore W3161026078C138885662 @default.
• W3161026078 hasConceptScore W3161026078C154945302 @default.
• W3161026078 hasConceptScore W3161026078C165696696 @default.
• W3161026078 hasConceptScore W3161026078C173801870 @default.
• W3161026078 hasConceptScore W3161026078C177264268 @default.
• W3161026078 hasConceptScore W3161026078C199360897 @default.
• W3161026078 hasConceptScore W3161026078C33923547 @default.
• W3161026078 hasConceptScore W3161026078C38652104 @default.
• W3161026078 hasConceptScore W3161026078C41008148 @default.
• W3161026078 hasConceptScore W3161026078C41895202 @default.
• W3161026078 hasConceptScore W3161026078C51823790 @default.
• W3161026078 hasConceptScore W3161026078C521332185 @default.
• W3161026078 hasLocation W31610260781 @default.
• W3161026078 hasOpenAccess W3161026078 @default.
• W3161026078 hasPrimaryLocation W31610260781 @default.
• W3161026078 hasRelatedWork W1541792914 @default.
• W3161026078 hasRelatedWork W1559162534 @default.
• W3161026078 hasRelatedWork W1767197940 @default.
• W3161026078 hasRelatedWork W1792660423 @default.
• W3161026078 hasRelatedWork W2015456385 @default.
• W3161026078 hasRelatedWork W2138677310 @default.
• W3161026078 hasRelatedWork W2276444382 @default.
• W3161026078 hasRelatedWork W2380031573 @default.
• W3161026078 hasRelatedWork W2413613665 @default.
• W3161026078 hasRelatedWork W2509717476 @default.
• W3161026078 hasRelatedWork W2544636806 @default.
• W3161026078 hasRelatedWork W2767267671 @default.
• W3161026078 hasRelatedWork W2767772787 @default.
• W3161026078 hasRelatedWork W2791733752 @default.
• W3161026078 hasRelatedWork W3105826019 @default.
• W3161026078 hasRelatedWork W3157651337 @default.
• W3161026078 hasRelatedWork W3173651597 @default.
• W3161026078 hasRelatedWork W3194015442 @default.
• W3161026078 hasRelatedWork W565860700 @default.
• W3161026078 hasRelatedWork W2560324307 @default.
• W3161026078 isParatext "false" @default.
• W3161026078 isRetracted "false" @default.
• W3161026078 magId "3161026078" @default.
• W3161026078 workType "article" @default.