Data and code from Price, Buckles, Van Leeuwen, & Riley (Explorations in Economic History)

These files include data and code used to create tables and figures for "Combining Family History and Machine Learning to Link Historical Records." This was published as an NBER Working Paper (#26227) and is in the review process at an academic journal. We are seeking permission to share our full dataset that we obtained from FamilySearch. We have included the truth set that we use to train our model along with the machine learning code we use to link records using this truth set. Our truth set includes histid pairs between the different census records, which can be combined with the restricted versions of the full-count census records that are distributed through the University of Minnesota.

本数据集包含用于构建《结合家族历史与机器学习以关联历史记录》一文中表格与图表的数据及代码。该文已作为NBER工作论文（编号#26227）发表，并正处于学术期刊的审稿阶段。我们正在寻求分享我们从FamilySearch获取的完整数据集的许可。数据集中包含了我们用于训练模型的真理集以及我们使用该真理集关联记录所采用的机器学习代码。我们的真理集包含了不同人口普查记录之间的histid对，这些对可以与通过明尼苏达大学分发的人口普查全记录的受限版本相结合。