---
title: DEFONEOS BFT Voter Cosine Similarity Cluster
description: Per-agent voter cohort cosine similarity analysis for the DEFONEOS BFT Council.
---
# DEFONEOS BFT Voter Cosine Similarity Cluster

This page presents a detailed analysis of the DEFONEOS BFT Council's voter cohorts based on their cosine similarity across various motion classes and voting patterns.

## Analysis Methodology
- **Cosine Similarity:** Measures the angle between voting vectors of individual agents, indicating the similarity of their voting behavior.
- **Clustering:** Agents are grouped into cohorts based on their cosine similarity scores.
- **Motion Classes:** Analysis performed across standard, red-line, constitutional, unanimous, and emergency motion classes.

## Cohort Breakdown
- **Alpha Cohort:** High participation, high similarity, core decision-makers.
- **Beta Cohort:** Moderate participation, moderate similarity, supporting roles.
- **Gamma Cohort:** Low participation, diverse voting, specialized input.
- **Delta Cohort:** Minimal participation, observer status.

## Key Metrics
- **Intra-cohort Similarity:** Average cosine similarity within each cohort.
- **Inter-cohort Dissimilarity:** Average cosine dissimilarity between cohorts.
- **Vote Event Heatmap:** Visual representation of voting activity across cohorts and motion types.

## Verification
- Cross-linked to `defoneos-bft-voter-participation-cohort.html` for raw participation data.
- Regularly updated with the latest BFT council vote data.

## Further Drill-down
Explore individual agent profiles and their historical voting records.