We investigate the dynamics of opinion maximization through voting strategies: Identify a limited number of seed users in a social network for a target campaigner, amidst competing campaigners, aiming to maximize a voting-based score for the target campaigner within a given time horizon. Diverging from the conventional assumption that users are being confined to binary states of being active or inactive, with the option to switch states frozen upon one-time activation, this work recognizes the dynamic nature of user preferences. Even with preferred opinions, users may exhibit openness to alternative viewpoints, and their preferences may evolve over time due to social influence. To fill the gap, we draw on models grounded in opinion formation and diffusion, explore diverse voting-based scores, and strategically apply them in various real-world scenarios to effectively capture a user’s evolving vote for multiple campaigners, in the context of a social network and within a given time horizon. Our problem is <InlineEquation ID="IEq1"> <InlineMediaObject> <ImageObject Color="BlackWhite" FileRef="778_2025_933_Article_IEq1.gif" Format="GIF" Height="14" Rendition="HTML" Resolution="72" Type="Linedraw" Width="29" /> </InlineMediaObject> <EquationSource Format="TEX">$\textbf{NP}$</EquationSource> <EquationSource Format="MATHML"><math> <mi mathvariant="bold">NP</mi> </math></EquationSource> </InlineEquation>-hard, does not have any <InlineEquation ID="IEq2"> <InlineMediaObject> <ImageObject Color="BlackWhite" FileRef="778_2025_933_Article_IEq2.gif" Format="GIF" Height="14" Rendition="HTML" Resolution="72" Type="Linedraw" Width="50" /> </InlineMediaObject> <EquationSource Format="TEX">$\textbf{PTAS}$</EquationSource> <EquationSource Format="MATHML"><math> <mi mathvariant="bold">PTAS</mi> </math></EquationSource> </InlineEquation> unless <InlineEquation ID="IEq3"> <InlineMediaObject> <ImageObject Color="BlackWhite" FileRef="778_2025_933_Article_IEq3.gif" Format="GIF" Height="14" Rendition="HTML" Resolution="72" Type="Linedraw" Width="14" /> </InlineMediaObject> <EquationSource Format="TEX">$\textbf{P}$</EquationSource> <EquationSource Format="MATHML"><math> <mi mathvariant="bold">P</mi> </math></EquationSource> </InlineEquation> <InlineEquation ID="IEq4"> <InlineMediaObject> <ImageObject Color="BlackWhite" FileRef="778_2025_933_Article_IEq4.gif" Format="GIF" Height="6" Rendition="HTML" Resolution="72" Type="Linedraw" Width="14" /> </InlineMediaObject> <EquationSource Format="TEX">$=$</EquationSource> <EquationSource Format="MATHML"><math> <mo>=</mo> </math></EquationSource> </InlineEquation> <InlineEquation ID="IEq5"> <InlineMediaObject> <ImageObject Color="BlackWhite" FileRef="778_2025_933_Article_IEq1.gif" Format="GIF" Height="14" Rendition="HTML" Resolution="72" Type="Linedraw" Width="29" /> </InlineMediaObject> <EquationSource Format="TEX">$\textbf{NP}$</EquationSource> <EquationSource Format="MATHML"><math> <mi mathvariant="bold">NP</mi> </math></EquationSource> </InlineEquation>, and is non-submodular for various scores. We design greedy seed selection algorithms with quality guarantees via sandwich approximation. To improve the efficiency, we develop random walk and sketch-based opinion computation, with quality guarantees. Empirical results validate our effectiveness, efficiency, and scalability.

Beyond influence: voting theory for opinion maximization

期刊论文
发表日期：

2025年7月2日

查看全文

Arkaprava Saha,
Xiangyu Ke,
Arijit Khan,
Laks V. S. Lakshmanan

摘要

We investigate the dynamics of opinion maximization through voting strategies: Identify a limited number of seed users in a social network for a target campaigner, amidst competing campaigners, aiming to maximize a voting-based score for the target campaigner within a given time horizon. Diverging from the conventional assumption that users are being confined to binary states of being active or inactive, with the option to switch states frozen upon one-time activation, this work recognizes the dynamic nature of user preferences. Even with preferred opinions, users may exhibit openness to alternative viewpoints, and their preferences may evolve over time due to social influence. To fill the gap, we draw on models grounded in opinion formation and diffusion, explore diverse voting-based scores, and strategically apply them in various real-world scenarios to effectively capture a user’s evolving vote for multiple campaigners, in the context of a social network and within a given time horizon. Our problem is $\textbf{NP}$ $NP$ -hard, does not have any $\textbf{PTAS}$ $PTAS$ unless $\textbf{P}$ $P$ $=$ $=$ $\textbf{NP}$ $NP$ , and is non-submodular for various scores. We design greedy seed selection algorithms with quality guarantees via sandwich approximation. To improve the efficiency, we develop random walk and sketch-based opinion computation, with quality guarantees. Empirical results validate our effectiveness, efficiency, and scalability.