High-Modularity Graph Partitioning Through NLP Techniques and Maximal Clique Enumeration

2026-02-27

AI Summary

This work addresses the classic combinatorial optimization problem of high-modularity graph partitioning by proposing Clique-TF-IDF, a novel approach that adapts the TF-IDF mechanism from natural language processing to graph partitioning. Specifically, it constructs a vertex-clique incidence matrix through maximal clique enumeration, drawing an analogy to the term-document representation in text analysis. This representation is then leveraged by machine learning–based clustering algorithms to produce high-quality partitions without requiring the number of clusters to be specified in advance. Evaluated on multiple benchmark graphs, the method achieves performance comparable to or better than state-of-the-art algorithms, offering a scalable and effective AI-driven paradigm for tackling combinatorial optimization problems.

Technology Category

• Machine Learning: Graph-based Machine Learning
• Search and Optimization: Combinatorial Optimization
• Constraint Satisfaction and Optimization: Distributed CSP/Optimization

Application Category

• Graph Algorithms and Modeling for the Web: Representation, reconstruction, and subgraph or motif discovery in Web-related graphs
• Web Mining and Content Analysis: Normalization, clustering, classification, and summarization of Web text
• Semantics and Knowledge: Methods, algorithms and applications for the development of semantic models, knowledge graphs and other forms of structured data models with machine-interpretable semantics

Abstract

Natural Language Processing (NLP) provides highly effective tools for interpreting and handling human language, offering a broad spectrum of applications. In this paper, we address a classic combinatorial problem -- finding graph partitions with high modularity -- by applying NLP techniques that compute term frequency and inverse document frequency (TF-IDF) alongside machine learning clustering algorithms. We present a new framework, called Clique-TF-IDF, designed for graph partitioning, a task that holds significant relevance across various network analysis contexts. This approach uses dense substructures of the graph, specifically maximal cliques, to represent each vertex in terms of the cliques it is part of, in a manner akin to term-document matrices. Experiments show that Clique-TF-IDF yields results that are comparable to or outperform the current state-of-the-art algorithms, whether or not the number of partitions is known in advance. Although this framework emphasizes on cliques and partitioning, it can be extended to devise AI-driven solutions for a variety of challenging combinatorial problems that can leverage efficiently enumerable substructures.

Problem

• graph partitioning
• modularity
• maximal clique
• combinatorial optimization
• network analysis

Innovation

• Clique-TF-IDF
• graph partitioning
• modularity optimization
• maximal clique enumeration
• TF-IDF