# Enhanced graph coevolution network for social network analysis using assimilation modified emotional algorithm

## TL;DR

This paper introduces an improved algorithm for social network analysis that better captures emotional features and network dynamics.

## Contribution

The AME algorithm enhances traditional label propagation by incorporating emotional transitions and multi-label encoding.

## Key findings

- The AME algorithm outperforms traditional LPA methods in accuracy and loss values.
- Incorporating emotional transitions improves dynamic network analysis performance.

## Abstract

This paper presents the Assimilation Modified Emotional (AME) algorithm, which is an enhanced version of the traditional label propagation algorithm (LPA) designed to address key challenges in social network analysis and emotional feature extraction. Traditional LPA methods, such as asynchronous label propagation and the Louvain algorithm, do not incorporate emotional representations and are often limited by local structural dependencies. The AME algorithm addresses these limitations by applying spectral algorithms, Markov chains, graph coarsening, and link prediction to simulate and optimize emotional transitions within the network. In addition, the AME algorithm enhances label representation through multi-label encoding, which allows for more accurate simulation of dynamic emotional states. Experimental results show that the AME algorithm achieves better performance than traditional LPA methods in terms of both accuracy and loss values. These findings indicate that the AME algorithm has strong potential for improving AI models used in social network analysis and emotional feature extraction.