Motivation:
Entity Matching isthe task of determining which records from different data sources describe thesame real-world entity. It is an important task for data integration and hasbeen the focus of many research works. A large number of entity matching/record linkage taskshas been made available for evaluating entity matching methods. However, thelack of fixed development and test splits as well as correspondence sets including both matching and non-matchingrecord pairs hinders the reproducibility and comparability of benchmarkexperiments. In an effort to enhance the reproducibility and comparability ofthe experiments, we complement existing entity matching benchmark tasks withfixed sets of non-matching pairs as well as fixed development and test splits.

Dataset Description:
An augmentedversion of the fodors-zagats restaurants dataset for benchmarking entity matching/record linkagemethods found at:
https://hpi.de/en/naumann/projects/data-integration-data-quality-and-data-cleansing/dude.html#c11471

The augmentedversion adds a fixed set of non-matching pairs to the original dataset. Inaddition, fixed splits for training, validation and testing as well as theircorresponding feature vectors are provided. The feature vectors are built usingdata type specific similarity metrics.

The dataset contains 533 records describing restaurants from fodors.com which are matched against 331 restaurants records from zagat.com. The gold standards have manual annotations for 112 matchingand 488 non-matching pairs. The total number of attributes used to decribethe product records are 5 while the attribute density is 100%.

The augmenteddataset enhances the reproducibility of matching methods and the comparabilityof matching results.
The dataset is partof the CompERBench repository which provides 21 complete benchmark tasksfor entity matching for public download:
http://data.dws.informatik.uni-mannheim.de/benchmarkmatchingtasks/index.html

Error Loading Remote Resources