Elevating Social Graphs: Trust-Based Relationship Formation through Tertiary Contacts

TL;DR

Most social network models stop at immediate neighbors, but real-world trust flows deeper. This paper proposes a novel growth algorithm that incorporates Neighbor of Neighbor of Initial Contact (NNIC) to simulate high-trust environments like marriage alliances. The result is a complex, scale-free network with a superior clustering coefficient of $c(k)\sim \ln k/k$, significantly outperforming previous 1st-degree referral models.

Background: The Trust Deficit in Network Science

While Barabási-Albert (BA) models and Small-World theories explain how networks become "scale-free" or "highly connected," they often miss the human element: Trust. In specific contexts—such as professional collaboration or community-based marriage alliances—connections aren't just random or based on popularity (preferential attachment); they are mediated through trusted intermediaries.

The author argues that prior models [19] utilized friend information but ignored the "friend-of-a-friend" (Tertiary Contact) potential, which is the cornerstone of trust-based referrals in many cultures, particularly in India.

Methodology: The Three-Tier Attachment Algorithm

The core innovation is the Network Growth Algorithm, which moves beyond the 2nd degree of separation to the 3rd degree.

The 3-Step Connection Process:

1. Random Attachment: A new vertex (family/person) picks $m_r$ initial contacts.
2. Secondary Contact (NIC): The vertex connects to $m_s$ neighbors of those initial contacts.
3. Tertiary Contact (NNIC): Crucially, the vertex then connects to $m_t$ neighbors of the secondary contacts.

Mathematical Intuition

By using the Mean Field Approximation, the author derives the rate equation for vertex degree ($k_i$): $\frac{\partial {\mathrm{k}}_{\mathrm{i}}}{\partial t}=\frac{1}{t}\left({\mathrm{m}}_{\mathrm{r}}+\frac{{\mathrm{m}}_{\mathrm{r}}{\mathrm{m}}_{\mathrm{s}}+2{\mathrm{m}}_{\mathrm{r}}{\mathrm{m}}_{\mathrm{s}}{\mathrm{m}}_{\mathrm{t}}}{2({\mathrm{m}}_{\mathrm{r}}+{\mathrm{m}}_{\mathrm{r}}{\mathrm{m}}_{\mathrm{s}}+2{\mathrm{m}}_{\mathrm{r}}{\mathrm{m}}_{\mathrm{s}}{\mathrm{m}}_{\mathrm{t}})}{\mathrm{k}}_{\mathrm{i}}\right)$

This formula suggests that the "preferential attachment" effect is amplified by the existence of secondary and tertiary tiers, allowing nodes that are well-connected deep in the network to be more easily discovered through trusted paths.

Experimental Results: High Clustering & Efficiency

The study simulated a network of 45 families to validate the growth dynamics. The introduction of NNIC contacts led to a dramatic increase in "Triangles" (mutual relationships), which is the bedrock of community structure.

Key Performance Metrics:

• Scale-Free Nature: The degree distribution follows $p(k)\sim k^{-\gamma }$, where $\gamma$ remains bounded at 3, preserving the robust architecture of a scale-free network.
• Clustering Leap: Previous models yielded $c(k)\sim 1/k$. This model achieves $c(k)\sim \ln k/k$. For large networks, this logarithmic improvement means the network is far more tightly knit than standard models would predict.

Critical Analysis & Future Outlook

The "NNIC" model successfully bridges the gap between abstract graph theory and sociological reality. By focusing on "tertiary contacts," it mathematically proves why "trust-based" searching is faster and yields better "matches" without diluting the reliability of the contacts.

Limitations:

• The model assumes a "homogeneous" community (e.g., same caste/region). Future iterations should explore inter-community interactions where trust may decay as one moves across different social clusters.
• The sample size of 45 families is useful for proof-of-concept but lacks the "large $N$" behavior analysis traditional in statistical physics.

Conclusion: This research is a significant step for designing specialized social platforms—be it for marriage, research collaboration, or high-stakes financial networking—where a simple "friend" suggestion isn't enough, but a "trusted referral of a referral" is everything.

Keywords: Scale-free Networks, Clustering Coefficient, Social Discovery, Trust-based Modeling, NNIC.