Abstract:
Identifying authors relies on their unique writing patterns, also known as stylometry. However, as each individual's stylometry may differ, it can be challenging to determine the true author, especially when multiple documents exist. This difficulty is compounded by similarities in writing styles, such as font and language, which can obscure the author's identity. To address this issue, machine learning techniques can be employed to identify human attitudes in written documents. By analyzing patterns in human behavior, it is possible to enhance privacy and security by detecting malicious users and malware programs. However, the use of behavior analysis may raise privacy concerns, particularly for individuals seeking to conceal their identity. Authorship attribution involves using stylometry techniques to identify the authors of multiple documents, and can aid in accurately identifying authors. In this research, we propose using stylometry to extract the number of programmers from a given database, and to analyze different datasets to determine whether a program's coding style remains consistent. This analysis can enhance the reliability and quality of programming, ultimately improving the overall efficiency of programming tasks.
Page(s):
28-35
DOI:
DOI not available
Published:
Journal: Technical Journal, Volume: 28, Issue: 1, Year: 2023
Keywords:
deep learning
,
Machine learning
,
programmer deanonymization
,
author identification
,
Stylometry