We propose a novel framework, called IncompFuse , that significantly improves the accuracy of existing methods for reconstructing aggregated historical data from inaccurate historical reports. IncompFuse supports efficient data reliability assessment using the incompatibility probability of historical reports. We provide a systematic approach to define this probability based on properties of the data and relationships between the reports. Our experimental study demonstrates high utility of the proposed framework. In particular, we were able to detect noisy historical reports with very high detection accuracy.

中文翻译：

---

IncompFuse：用于历史信息与不准确数据源融合的逻辑框架

我们提出了一个名为 IncompFuse 的新框架，它显着提高了从不准确的历史报告中重建聚合历史数据的现有方法的准确性。IncompFuse 支持使用历史报告的不兼容概率进行有效的数据可靠性评估。我们提供了一种系统的方法来根据数据的属性和报告之间的关系来定义这种概率。我们的实验研究证明了所提出框架的高实用性。特别是，我们能够以非常高的检测精度检测嘈杂的历史报告。