TY - JOUR A1 - Koumarelas, Ioannis A1 - Papenbrock, Thorsten A1 - Naumann, Felix T1 - MDedup BT - duplicate detection with matching dependencies JF - Proceedings of the VLDB Endowment N2 - Duplicate detection is an integral part of data cleaning and serves to identify multiple representations of same real-world entities in (relational) datasets. Existing duplicate detection approaches are effective, but they are also hard to parameterize or require a lot of pre-labeled training data. Both parameterization and pre-labeling are at least domain-specific if not dataset-specific, which is a problem if a new dataset needs to be cleaned. For this reason, we propose a novel, rule-based and fully automatic duplicate detection approach that is based on matching dependencies (MDs). Our system uses automatically discovered MDs, various dataset features, and known gold standards to train a model that selects MDs as duplicate detection rules. Once trained, the model can select useful MDs for duplicate detection on any new dataset. To increase the generally low recall of MD-based data cleaning approaches, we propose an additional boosting step. Our experiments show that this approach reaches up to 94% F-measure and 100% precision on our evaluation datasets, which are good numbers considering that the system does not require domain or target data-specific configuration. Y1 - 2020 U6 - https://doi.org/10.14778/3377369.3377379 SN - 2150-8097 VL - 13 IS - 5 SP - 712 EP - 725 PB - Association for Computing Machinery CY - New York ER - TY - JOUR A1 - Koumarelas, Ioannis A1 - Jiang, Lan A1 - Naumann, Felix T1 - Data preparation for duplicate detection JF - Journal of data and information quality : (JDIQ) N2 - Data errors represent a major issue in most application workflows. Before any important task can take place, a certain data quality has to be guaranteed by eliminating a number of different errors that may appear in data. Typically, most of these errors are fixed with data preparation methods, such as whitespace removal. However, the particular error of duplicate records, where multiple records refer to the same entity, is usually eliminated independently with specialized techniques. Our work is the first to bring these two areas together by applying data preparation operations under a systematic approach prior to performing duplicate detection.
Our process workflow can be summarized as follows: It begins with the user providing as input a sample of the gold standard, the actual dataset, and optionally some constraints to domain-specific data preparations, such as address normalization. The preparation selection operates in two consecutive phases. First, to vastly reduce the search space of ineffective data preparations, decisions are made based on the improvement or worsening of pair similarities. Second, using the remaining data preparations an iterative leave-one-out classification process removes preparations one by one and determines the redundant preparations based on the achieved area under the precision-recall curve (AUC-PR). Using this workflow, we manage to improve the results of duplicate detection up to 19% in AUC-PR. KW - data preparation KW - data wrangling KW - record linkage KW - duplicate detection KW - similarity measures Y1 - 2020 U6 - https://doi.org/10.1145/3377878 SN - 1936-1955 SN - 1936-1963 VL - 12 IS - 3 PB - Association for Computing Machinery CY - New York ER - TY - BOOK A1 - Draisbach, Uwe A1 - Naumann, Felix A1 - Szott, Sascha A1 - Wonneberg, Oliver T1 - Adaptive windows for duplicate detection N2 - Duplicate detection is the task of identifying all groups of records within a data set that represent the same real-world entity, respectively. This task is difficult, because (i) representations might differ slightly, so some similarity measure must be defined to compare pairs of records and (ii) data sets might have a high volume making a pair-wise comparison of all records infeasible. To tackle the second problem, many algorithms have been suggested that partition the data set and compare all record pairs only within each partition. One well-known such approach is the Sorted Neighborhood Method (SNM), which sorts the data according to some key and then advances a window over the data comparing only records that appear within the same window. We propose several variations of SNM that have in common a varying window size and advancement. The general intuition of such adaptive windows is that there might be regions of high similarity suggesting a larger window size and regions of lower similarity suggesting a smaller window size. We propose and thoroughly evaluate several adaption strategies, some of which are provably better than the original SNM in terms of efficiency (same results with fewer comparisons). N2 - Duplikaterkennung beschreibt das Auffinden von mehreren Datensätzen, die das gleiche Realwelt-Objekt repräsentieren. Diese Aufgabe ist nicht trivial, da sich (i) die Datensätze geringfügig unterscheiden können, so dass Ähnlichkeitsmaße für einen paarweisen Vergleich benötigt werden, und (ii) aufgrund der Datenmenge ein vollständiger, paarweiser Vergleich nicht möglich ist. Zur Lösung des zweiten Problems existieren verschiedene Algorithmen, die die Datenmenge partitionieren und nur noch innerhalb der Partitionen Vergleiche durchführen. Einer dieser Algorithmen ist die Sorted-Neighborhood-Methode (SNM), welche Daten anhand eines Schlüssels sortiert und dann ein Fenster über die sortierten Daten schiebt. Vergleiche werden nur innerhalb dieses Fensters durchgeführt. Wir beschreiben verschiedene Variationen der Sorted-Neighborhood-Methode, die auf variierenden Fenstergrößen basieren. Diese Ansätze basieren auf der Intuition, dass Bereiche mit größerer und geringerer Ähnlichkeiten innerhalb der sortierten Datensätze existieren, für die entsprechend größere bzw. kleinere Fenstergrößen sinnvoll sind. Wir beschreiben und evaluieren verschiedene Adaptierungs-Strategien, von denen nachweislich einige bezüglich Effizienz besser sind als die originale Sorted-Neighborhood-Methode (gleiches Ergebnis bei weniger Vergleichen). T3 - Technische Berichte des Hasso-Plattner-Instituts für Digital Engineering an der Universität Potsdam - 49 KW - Informationssysteme KW - Datenqualität KW - Datenintegration KW - Duplikaterkennung KW - Duplicate Detection KW - Data Quality KW - Data Integration KW - Information Systems Y1 - 2012 U6 - http://nbn-resolving.de/urn/resolver.pl?urn:nbn:de:kobv:517-opus-53007 SN - 978-3-86956-143-1 SN - 1613-5652 SN - 2191-1665 PB - Universitätsverlag Potsdam CY - Potsdam ER - TY - JOUR A1 - Koumarelas, Ioannis A1 - Kroschk, Axel A1 - Mosley, Clifford A1 - Naumann, Felix T1 - Experience: Enhancing address matching with geocoding and similarity measure selection JF - Journal of Data and Information Quality N2 - Given a query record, record matching is the problem of finding database records that represent the same real-world object. In the easiest scenario, a database record is completely identical to the query. However, in most cases, problems do arise, for instance, as a result of data errors or data integrated from multiple sources or received from restrictive form fields. These problems are usually difficult, because they require a variety of actions, including field segmentation, decoding of values, and similarity comparisons, each requiring some domain knowledge. In this article, we study the problem of matching records that contain address information, including attributes such as Street-address and City. To facilitate this matching process, we propose a domain-specific procedure to, first, enrich each record with a more complete representation of the address information through geocoding and reverse-geocoding and, second, to select the best similarity measure per each address attribute that will finally help the classifier to achieve the best f-measure. We report on our experience in selecting geocoding services and discovering similarity measures for a concrete but common industry use-case. KW - Address matching KW - record linkage KW - duplicate detection KW - similarity measures KW - conditional functional dependencies KW - address normalization KW - address parsing KW - geocoding KW - geographic information systems KW - random forest Y1 - 2018 U6 - https://doi.org/10.1145/3232852 SN - 1936-1955 VL - 10 IS - 2 SP - 1 EP - 16 PB - Association for Computing Machinery CY - New York ER - TY - JOUR A1 - Hameed, Mazhar A1 - Naumann, Felix T1 - Data Preparation BT - a survey of commercial tools JF - SIGMOD record N2 - Raw data are often messy: they follow different encodings, records are not well structured, values do not adhere to patterns, etc. Such data are in general not fit to be ingested by downstream applications, such as data analytics tools, or even by data management systems. The act of obtaining information from raw data relies on some data preparation process. Data preparation is integral to advanced data analysis and data management, not only for data science but for any data-driven applications. Existing data preparation tools are operational and useful, but there is still room for improvement and optimization. With increasing data volume and its messy nature, the demand for prepared data increases day by day.
To cater to this demand, companies and researchers are developing techniques and tools for data preparation. To better understand the available data preparation systems, we have conducted a survey to investigate (1) prominent data preparation tools, (2) distinctive tool features, (3) the need for preliminary data processing even for these tools and, (4) features and abilities that are still lacking. We conclude with an argument in support of automatic and intelligent data preparation beyond traditional and simplistic techniques. KW - data quality KW - data cleaning KW - data wrangling Y1 - 2020 U6 - https://doi.org/10.1145/3444831.3444835 SN - 0163-5808 SN - 1943-5835 VL - 49 IS - 3 SP - 18 EP - 29 PB - Association for Computing Machinery CY - New York ER - TY - JOUR A1 - Schirmer, Philipp A1 - Papenbrock, Thorsten A1 - Koumarelas, Ioannis A1 - Naumann, Felix T1 - Efficient discovery of matching dependencies JF - ACM transactions on database systems : TODS N2 - Matching dependencies (MDs) are data profiling results that are often used for data integration, data cleaning, and entity matching. They are a generalization of functional dependencies (FDs) matching similar rather than same elements. As their discovery is very difficult, existing profiling algorithms find either only small subsets of all MDs or their scope is limited to only small datasets. We focus on the efficient discovery of all interesting MDs in real-world datasets. For this purpose, we propose HyMD, a novel MD discovery algorithm that finds all minimal, non-trivial MDs within given similarity boundaries. The algorithm extracts the exact similarity thresholds for the individual MDs from the data instead of using predefined similarity thresholds. For this reason, it is the first approach to solve the MD discovery problem in an exact and truly complete way. If needed, the algorithm can, however, enforce certain properties on the reported MDs, such as disjointness and minimum support, to focus the discovery on such results that are actually required by downstream use cases. HyMD is technically a hybrid approach that combines the two most popular dependency discovery strategies in related work: lattice traversal and inference from record pairs. Despite the additional effort of finding exact similarity thresholds for all MD candidates, the algorithm is still able to efficiently process large datasets, e.g., datasets larger than 3 GB. KW - matching dependencies KW - functional dependencies KW - dependency discovery KW - data profiling KW - data matching KW - entity resolution KW - similarity measures Y1 - 2020 U6 - https://doi.org/10.1145/3392778 SN - 0362-5915 SN - 1557-4644 VL - 45 IS - 3 PB - Association for Computing Machinery CY - New York ER - TY - JOUR A1 - Hacker, Philipp A1 - Krestel, Ralf A1 - Grundmann, Stefan A1 - Naumann, Felix T1 - Explainable AI under contract and tort law BT - legal incentives and technical challenges JF - Artificial intelligence and law N2 - This paper shows that the law, in subtle ways, may set hitherto unrecognized incentives for the adoption of explainable machine learning applications. In doing so, we make two novel contributions. First, on the legal side, we show that to avoid liability, professional actors, such as doctors and managers, may soon be legally compelled to use explainable ML models. We argue that the importance of explainability reaches far beyond data protection law, and crucially influences questions of contractual and tort liability for the use of ML models. To this effect, we conduct two legal case studies, in medical and corporate merger applications of ML. As a second contribution, we discuss the (legally required) trade-off between accuracy and explainability and demonstrate the effect in a technical case study in the context of spam classification. KW - explainability KW - explainable AI KW - interpretable machine learning KW - contract KW - law KW - tort law KW - explainability-accuracy trade-off KW - medical malpractice KW - corporate takeovers Y1 - 2020 U6 - https://doi.org/10.1007/s10506-020-09260-6 SN - 0924-8463 SN - 1572-8382 VL - 28 IS - 4 SP - 415 EP - 439 PB - Springer CY - Dordrecht ER - TY - JOUR A1 - Draisbach, Uwe A1 - Christen, Peter A1 - Naumann, Felix T1 - Transforming pairwise duplicates to entity clusters for high-quality duplicate detection JF - ACM Journal of Data and Information Quality N2 - Duplicate detection algorithms produce clusters of database records, each cluster representing a single real-world entity. As most of these algorithms use pairwise comparisons, the resulting (transitive) clusters can be inconsistent: Not all records within a cluster are sufficiently similar to be classified as duplicate. Thus, one of many subsequent clustering algorithms can further improve the result.
We explain in detail, compare, and evaluate many of these algorithms and introduce three new clustering algorithms in the specific context of duplicate detection. Two of our three new algorithms use the structure of the input graph to create consistent clusters. Our third algorithm, and many other clustering algorithms, focus on the edge weights, instead. For evaluation, in contrast to related work, we experiment on true real-world datasets, and in addition examine in great detail various pair-selection strategies used in practice. While no overall winner emerges, we are able to identify best approaches for different situations. In scenarios with larger clusters, our proposed algorithm, Extended Maximum Clique Clustering (EMCC), and Markov Clustering show the best results. EMCC especially outperforms Markov Clustering regarding the precision of the results and additionally has the advantage that it can also be used in scenarios where edge weights are not available. KW - Record linkage KW - data matching KW - entity resolution KW - deduplication KW - clustering Y1 - 2019 U6 - https://doi.org/10.1145/3352591 SN - 1936-1955 SN - 1936-1963 VL - 12 IS - 1 SP - 1 EP - 30 PB - Association for Computing Machinery CY - New York ER - TY - JOUR A1 - Momtazi, Saeedeh A1 - Naumann, Felix T1 - Topic modeling for expert finding using latent Dirichlet allocation JF - Wiley interdisciplinary reviews : Data mining and knowledge discovery N2 - The task of expert finding is to rank the experts in the search space given a field of expertise as an input query. In this paper, we propose a topic modeling approach for this task. The proposed model uses latent Dirichlet allocation (LDA) to induce probabilistic topics. In the first step of our algorithm, the main topics of a document collection are extracted using LDA. The extracted topics present the connection between expert candidates and user queries. In the second step, the topics are used as a bridge to find the probability of selecting each candidate for a given query. The candidates are then ranked based on these probabilities. The experimental results on the Text REtrieval Conference (TREC) Enterprise track for 2005 and 2006 show that the proposed topic-based approach outperforms the state-of-the-art profile- and document-based models, which use information retrieval methods to rank experts. Moreover, we present the superiority of the proposed topic-based approach to the improved document-based expert finding systems, which consider additional information such as local context, candidate prior, and query expansion. Y1 - 2013 U6 - https://doi.org/10.1002/widm.1102 SN - 1942-4787 VL - 3 IS - 5 SP - 346 EP - 353 PB - Wiley CY - San Fransisco ER - TY - BOOK A1 - Albrecht, Alexander A1 - Naumann, Felix T1 - Understanding cryptic schemata in large extract-transform-load systems N2 - Extract-Transform-Load (ETL) tools are used for the creation, maintenance, and evolution of data warehouses, data marts, and operational data stores. ETL workflows populate those systems with data from various data sources by specifying and executing a DAG of transformations. Over time, hundreds of individual workflows evolve as new sources and new requirements are integrated into the system. The maintenance and evolution of large-scale ETL systems requires much time and manual effort. A key problem is to understand the meaning of unfamiliar attribute labels in source and target databases and ETL transformations. Hard-to-understand attribute labels lead to frustration and time spent to develop and understand ETL workflows. We present a schema decryption technique to support ETL developers in understanding cryptic schemata of sources, targets, and ETL transformations. For a given ETL system, our recommender-like approach leverages the large number of mapped attribute labels in existing ETL workflows to produce good and meaningful decryptions. In this way we are able to decrypt attribute labels consisting of a number of unfamiliar few-letter abbreviations, such as UNP_PEN_INT, which we can decrypt to UNPAID_PENALTY_INTEREST. We evaluate our schema decryption approach on three real-world repositories of ETL workflows and show that our approach is able to suggest high-quality decryptions for cryptic attribute labels in a given schema. N2 - Extract-Transform-Load (ETL) Tools werden häufig beim Erstellen, der Wartung und der Weiterentwicklung von Data Warehouses, Data Marts und operationalen Datenbanken verwendet. ETL Workflows befüllen diese Systeme mit Daten aus vielen unterschiedlichen Quellsystemen. Ein ETL Workflow besteht aus mehreren Transformationsschritten, die einen DAG-strukturierter Graphen bilden. Mit der Zeit entstehen hunderte individueller ETL Workflows, da neue Datenquellen integriert oder neue Anforderungen umgesetzt werden müssen. Die Wartung und Weiterentwicklung von großen ETL Systemen benötigt viel Zeit und manuelle Arbeit. Ein zentrales Problem ist dabei das Verständnis unbekannter Attributnamen in Quell- und Zieldatenbanken und ETL Transformationen. Schwer verständliche Attributnamen führen zu Frustration und hohen Zeitaufwänden bei der Entwicklung und dem Verständnis von ETL Workflows. Wir präsentieren eine Schema Decryption Technik, die ETL Entwicklern das Verständnis kryptischer Schemata in Quell- und Zieldatenbanken und ETL Transformationen erleichtert. Unser Ansatz berücksichtigt für ein gegebenes ETL System die Vielzahl verknüpfter Attributnamen in den existierenden ETL Workflows. So werden gute und aussagekräftige "Decryptions" gefunden und wir sind in der Lage Attributnamen, die aus unbekannten Abkürzungen bestehen, zu "decrypten". So wird z.B. für den Attributenamen UNP_PEN_INT als Decryption UNPAIN_PENALTY_INTEREST vorgeschlagen. Unser Schema Decryption Ansatz wurde für drei ETL-Repositories evaluiert und es zeigte sich, dass unser Ansatz qualitativ hochwertige Decryptions für kryptische Attributnamen vorschlägt. T3 - Technische Berichte des Hasso-Plattner-Instituts für Digital Engineering an der Universität Potsdam - 60 KW - Extract-Transform-Load (ETL) KW - Data Warehouse KW - Datenintegration KW - Extract-Transform-Load (ETL) KW - Data Warehouse KW - Data Integration Y1 - 2012 U6 - http://nbn-resolving.de/urn/resolver.pl?urn:nbn:de:kobv:517-opus-61257 SN - 978-3-86956-201-8 PB - Universitätsverlag Potsdam CY - Potsdam ER -