Description

This artifact contains the code, data, and additional information for reproducing the results found in the ESEC/FSE 2022 paper entitled “23 Shades of Self-Admitted Technical Debt: An Empirical Study on Machine Learning Software”. This study analyzes the occurrence of self-admitted technical debt (SATD) in a dataset consisting of 2,641 open-source machine learning repositories. The artifact contains the Boa scripts ran to acquire the comment data, as well as the Python scripts which were used to filter the dataset into 68,820 SATD comments. During the dataset creation, a sample was taken for two authors to independently label before settling disagreements in discussion of a moderator. The authors’ labels and the agreed upon labels are included within the artifact.

Provenance

This study analyzes the occurrence of self-admitted technical debt (SATD) in a dataset consisting of 2,641 open-source machine learning repositories. The artifact contains the Boa scripts ran to acquire the comment data, as well as the Python scripts which were used to filter the dataset into 68,820 SATD comments. During the dataset creation, a sample was taken for two authors to independently label before settling disagreements in discussion of a moderator. The authors’ labels and the agreed upon labels are included within the artifact.

License