Dataset Information

From Leiden to Tel-Aviv University (TAU): exploring clustering solutions via a genetic algorithm.

ABSTRACT: Graph clustering is a fundamental problem in machine learning with numerous applications in data science. State-of-the-art approaches to the problem, Louvain and Leiden, aim at optimizing the modularity function. However, their greedy nature leads to fast convergence to sub-optimal solutions. Here, we design a new approach to graph clustering, Tel-Aviv University (TAU), that efficiently explores the solution space using a genetic algorithm. We benchmark TAU on synthetic and real data sets and show its superiority over previous methods both in terms of the modularity of the computed solution and its similarity to a ground-truth partition when such exists. TAU is available at https://github.com/GalGilad/TAU.

SUBMITTER: Gilad G

PROVIDER: S-EPMC10244004 | biostudies-literature | 2023 Jun

REPOSITORIES: biostudies-literature

ACCESS DATA

Publications

From Leiden to Tel-Aviv University (TAU): exploring clustering solutions via a genetic algorithm.

Gilad Gal G Sharan Roded R

PNAS nexus 20230601 6

Graph clustering is a fundamental problem in machine learning with numerous applications in data science. State-of-the-art approaches to the problem, Louvain and Leiden, aim at optimizing the modularity function. However, their greedy nature leads to fast convergence to sub-optimal solutions. Here, we design a new approach to graph clustering, Tel-Aviv University (TAU), that efficiently explores the solution space using a genetic algorithm. We benchmark TAU on synthetic and real data sets and sh ...[more]

PMID: 37287709

Similar Datasets

Project description:The purpose of this pilot cross-national study was to uncover similarities and differences in three areas that might affect the development of community-based programs targeting child maltreatment: behaviors considered to be maltreatment, perceived contributors to maltreatment, and whether the government or neighbors can do anything about maltreatment. Data were obtained from two neighborhood-based, cross-sectional surveys of adult caregivers of minors: one in Cleveland, USA, the other in Tel Aviv, Israel. The sample consisted of a total of 120 caregivers, in each city 20 residing in a low-SES neighborhood, 20 in a medium-SES neighborhood, and 20 in an elevated-SES neighborhood. Participants were asked (a) to provide three examples of behaviors they considered to be child abuse, (b) to rate the degree to which each of 13 factors contribute to child maltreatment, and (c) to rate the degree to which they agreed with a range of attitudes about maltreatment. The same coding scheme was used in both sites. Logistic regression analyses assessed city differences in dichotomous outcomes, while linear regression analyses assessed city differences in ratings of continuous outcomes. Analyses adjusted for individual and neighborhood characteristics, and accounted for residential clustering in neighborhoods. Primary results indicated that residence in Tel Aviv was associated with greater odds of citing emotional/psychological abuse compared to Cleveland residents. Also compared to Cleveland residents, Tel Aviv residents (a) viewed family structure, family values, religion, child-raising knowledge, and personal history of maltreatment as contributing less to maltreatment, (b) were less likely to agree that anyone could abuse a child or that spanking is necessary, and (c) had substantially greater odds of endorsing the government's ability to address child maltreatment. Concerning study implications, this investigation demonstrated the importance of context in shaping constructions of child maltreatment and the need for caution in replicating interventions without due consideration of potential differences in context, policy, and public opinion.

Dataset Information

From Leiden to Tel-Aviv University (TAU): exploring clustering solutions via a genetic algorithm.

Publications

From Leiden to Tel-Aviv University (TAU): exploring clustering solutions via a genetic algorithm.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets