mosaic-benchmark 0.0.2

A package to create modular link streams used for testing dynamic community detection algoritms

Mosaic Benchmark: Synthetic Modular link streams for testing dynamic community detection algorithms

Temporal networks offer valuable insights into dynamic complex systems, capturing the evolving nature of social, biological, and technological interactions. Community structure is a critical feature of real networks, revealing the internal organization of nodes. Dynamic community detection algorithms uncover strongly connected node groups, unveiling hidden temporal patterns and community dynamics in temporal networks.

However, evaluating the performance of these algorithms remains a challenge. A well-established method is to use tests that rely on synthetic graphs. Yet, this approach does not suit temporal networks with instantaneous edges and continuous time domains, known as link streams. To address this gap, we propose a novel benchmark comprising predefined communities that simulate synthetic modular link streams.

Mosaic is a meta-library for creating modular link streams for testing dynamic community detection algorithms in complex temporal networks: it creates communities, visualises them and exports the network to csv files.

Citation

If you use our algorithm, please cite the following works:

Yasaman Asgari Paper

Installation

To install the package, download (or clone) the current project and copy the demon folder in the root of your application.

Alternatively, use pip:

sudo pip install mosaic

Execution

The algorithm can be used as a standalone program and integrated into Python scripts.

Link stream parameters

Name	Type	Description
number_of_nodes	Integer	Number of nodes
starting_time	Float	Starting time for link stream's time domain
ending_time	Float	Ending time for link stream's time domain

Community parameters

Name	Type	Description
nodes	List	list of Integer indices inside a community
starting_time	Float	Starting time for community
ending_time	Float	Ending time for community

Scenario parameters

Name	Type	Description
$\alpha$	Float	[0.5,1): defines the internal density coefficient
$\beta$	Float	[0,1]: related to community identifialbility
$\eta$	Float	[0,1]: rewiring noise probability

Output

Mosaic Benchmark can export two different types of files.

• graph-*.csv: Edgelist representation of the generated graph. One line for each edge in link stream.
• communities-*.txt: community description. One file for stable iteration.

The syntax of each class of output files is the following:

Communities

A community per line descibed as:

community_id	[node1, node2, node3, ..., nodeN] stating_time  ending_time

Interactions

One edge per line with the syntax:

time node1 node2

Where:

• time identify the time in which the interaction occurs
• node1 and node2 are interaction endpoints

Example:

1.5	2	4
3	4	8

Dependencies

Mosaic is written in Python and requires the following package to run:

• python>=3.8
• Pandas
• tqdm
• Numpy
• Matplotlib
• itertools

Tutorial

Check out the official tutorial to get started!