Authorship attribution with traditional methods and complex networks

• Grant number: 15/23803-7
• Support Opportunities: Scholarships abroad - Research Internship - Master's degree
• Start date: March 01, 2016
• End date: August 31, 2016
• Field of knowledge: Physical Sciences and Mathematics - Computer Science - Computing Methodologies and Techniques
• Principal Investigator: Diego Raphael Amancio
• Grantee: Vanessa Queiroz Marinho
• Supervisor: Graeme Hirst
• Host Institution: Instituto de Ciências Matemáticas e de Computação (ICMC). Universidade de São Paulo (USP). São Carlos , SP, Brazil
• Institution abroad: University of Toronto (U of T), Canada
• Associated to the scholarship: 15/05676-8 - Development of new models for authorship recognition using complex networks, BP.MS

Abstract

Concepts and methods of complex networks have proven useful to probe several real systems of very distinct nature. The discovery that methods from complex networks can be used to analyse texts in their different complexity levels has allowed the study of naturallanguage processing (NLP) tasks from a new perspective. Examples of tasks studied via topological analysis of networks are keyword identification, automatic extractive summarization and authorship attribution. The latter task, which is the focus of this project, has been studied with some success by representing texts as words adjacency networks. Even though networked representations have been applied to study the authorship recognition problem, such approaches have not outperformed other traditional models relying upon statistical paradigms. Because network models are able to grasp textual patterns that can not be with traditional statistical models, we intend to devise hybrid systems that combine both traditional NLP techniques with properties provided by the topological analysis of complex networks. By combining such distinct paradigms in a complementary way, we aim to improve the performance of textual stylistic characterization and authorship attribution systems. We are bold to predict that such combination shall probably improve the performance of related applications, such as the analysis of stylistic inconsistencies, scientific frauds and plagiarism. (AU)

Scientific publications

(References retrieved automatically from Web of Science and SciELO through information on FAPESP grants and their corresponding numbers as mentioned in the publications by the authors)

MARINHO, VANESSA QUEIROZ; HIRST, GRAEME; AMANCIO, DIEGO RAPHAEL. Labelled network subgraphs reveal stylistic subtleties in written texts. JOURNAL OF COMPLEX NETWORKS, v. 6, n. 4, p. 620-638, AUG 2018. (15/05676-8, 14/20830-0, 15/23803-7, 16/19069-9)

MARINHO, VANESSA QUEIROZ; HIRST, GRAEME; AMANCIO, DIEGO RAPHAEL; IEEE. Authorship attribution via network motifs identification. PROCEEDINGS OF 2016 5TH BRAZILIAN CONFERENCE ON INTELLIGENT SYSTEMS (BRACIS 2016), v. N/A, p. 6-pg., 2016-01-01. (15/05676-8, 14/20830-0, 15/23803-7)