Higher-order Interactions Disturb Community Detection in Complex Networks

This repository contains code and data processing steps used in the study "Higher-order interactions disturb community detection in complex networks." The study focuses on the impact of higher-order interactions on community detection within complex networks.

Network Data Preprocessing

1. bip2one.m - Converts bipartite networks to unilayer networks.
2. high_order_nw.m - Transforms bipartite networks into unilayer and multilayer networks.

Model Construction

1. build_modle.m - Constructs the cross-community cooperation model. In this model, higher-order interactions are more likely to occur across communities.
2. build_modle_rand.m - Establishes a random model where interactions occur freely, independent of the orders.

Analysis of Imbalance in Number of Interactions and Edges

1. statistic.m - Demonstrates the imbalance in the number of interactions (i.e., papers) and edges per order.

High-order Interference in Community Partitioning in the Model

1. high_order_nosie.m - Script for plotting Figure 3, revealing the interference of high-order interactions in community detection within the model.

Higher Probability of Cross-Community Cooperation in Higher Orders

1. order_weight.m - Compares empirical networks and models to validate cross-community behaviors in high-order interactions in real systems.

Removing High-order Information Benefits Community Partitioning

1. pacs_deal_data.m - Processes the APS dataset, correlating authors with PACS codes of papers.
2. pacs_sim_big_author.m - Validates the consistency and effectiveness of field identification before and after removing high-order interactions.

Data Availability

Due to the size limitations of the GitHub repository, the complete dataset could not be uploaded here.

Original Data

The APS data are available upon request. Please submit your request to the American Physical Society (APS) at https://journals.aps.org/datasets.

Processed Data

For access to the processed data used in this study, please send an email inquiry to [email protected].

---

For details on the methods used in these scripts, as well as the analysis of results and conclusions, please refer to the corresponding sections of the research paper.