Identity Resolution Dataset
收藏资源简介:
Identity Resolution Dataset是由田纳西大学诺克斯维尔分校的研究团队创建,旨在解决开源软件项目中作者身份识别问题。该数据集包含从20亿Git提交中提取的3800万作者ID,通过机器学习模型预测哪些ID属于同一开发者。数据集大小为14,861,538条记录,每条记录代表一个具有别名的作者ID。创建过程中,研究团队首先使用启发式方法将作者ID分块,然后应用机器学习模型进行预测。该数据集可用于提高开源软件生态系统中开发者行为模型的准确性,并快速解决新的作者ID识别问题。
The Identity Resolution Dataset was developed by a research team from the University of Tennessee, Knoxville, with the goal of solving the author identification issue in open-source software projects. This dataset includes 38 million author IDs extracted from 2 billion Git commits, and is designed to predict whether different IDs belong to the same developer through machine learning models. It consists of 14,861,538 records in total, with each record representing an author ID with aliases. During the dataset construction process, the research team first employed heuristic methods to partition author IDs into blocks, and then applied machine learning models to perform the prediction task. This dataset can be utilized to enhance the accuracy of developer behavior models within the open-source software ecosystem, and to rapidly resolve new author ID identification problems.

- 1A Dataset and an Approach for Identity Resolution of 38 Million Author IDs extracted from 2B Git Commits田纳西大学诺克斯维尔分校 · 2020年



