2019

• 
  Dürsch, F., Stebner, A., Windheuser, F., Fischer, M., Friedrich, T., Strelow, N., Bleifuß, T., Harmouch, H., Jiang, L., Papenbrock, T., Naumann, F.: Inclusion Dependency Discovery: An Experimental Evaluation of Thirteen Algorithms.Proceedings of the International Conference on Information and Knowledge Management (CIKM). pp. 219–228 (2019).
  [ BibTeX ] [ Download ] [ Details ]
   
  @inproceedings{dursch2019inclusion, author = {Dürsch, Falco and Stebner, Axel and Windheuser, Fabian and Fischer, Maxi and Friedrich, Tim and Strelow, Nils and Bleifuß, Tobias and Harmouch, Hazar and Jiang, Lan and Papenbrock, Thorsten and Naumann, Felix}, booktitle = {Proceedings of the International Conference on Information and Knowledge Management (CIKM)}, keywords = {isg}, pages = {219–228}, title = {Inclusion Dependency Discovery: An Experimental Evaluation of Thirteen Algorithms}, year = 2019 }
  Further Information

2018

• 
  Berti-Equille, L., Harmouch, H., Naumann, F., Novelli, N., Thirumuruganathan, S.: Discovery of Genuine Functional Dependencies from Relational Data with Missing Values.Proceedings of the VLDB Endowment (PVLDB). pp. 880-892 (2018).
  [ Abstract ] [ BibTeX ] [ URL ] [ Download ] [ Details ]
   
  Functional dependencies (FDs) play an important role in maintaining data quality. They can be used to enforce data consistency and to guide repairs over a database. In this work, we investigate the problem of missing values and its impact on FD discovery. When using existing FD discovery algorithms, some genuine FDs could not be detected precisely due to missing values or some non-genuine FDs can be discovered even though they are caused by missing values with a certain NULL semantics. We define a notion of genuineness and propose algorithms to compute the genuineness score of a discovered FD. This can be used to identify the genuine FDs among the set of all valid dependencies that hold on the data. We evaluate the quality of our method over various real-world and semi-synthetic datasets with extensive experiments. The results show that our method performs well for relatively large FD sets and is able to accurately capture genuine FDs.
  @inproceedings{bertiequille2018discovery, abstract = {Functional dependencies (FDs) play an important role in maintaining data quality. They can be used to enforce data consistency and to guide repairs over a database. In this work, we investigate the problem of missing values and its impact on FD discovery. When using existing FD discovery algorithms, some genuine FDs could not be detected precisely due to missing values or some non-genuine FDs can be discovered even though they are caused by missing values with a certain NULL semantics. We define a notion of genuineness and propose algorithms to compute the genuineness score of a discovered FD. This can be used to identify the genuine FDs among the set of all valid dependencies that hold on the data. We evaluate the quality of our method over various real-world and semi-synthetic datasets with extensive experiments. The results show that our method performs well for relatively large FD sets and is able to accurately capture genuine FDs.}, author = {Berti-Equille, Laure and Harmouch, Hazar and Naumann, Felix and Novelli, Noel and Thirumuruganathan, Saravanan}, booktitle = {Proceedings of the VLDB Endowment (PVLDB)}, keywords = {null profiling discovery missing isg functional_dependencies}, month = 4, number = 8, pages = {880-892}, title = {Discovery of Genuine Functional Dependencies from Relational Data with Missing Values}, volume = 11, year = 2018 }
  Further Information

  URNhttp://dx.doi.org/10.14778/3204028.3204032

  AbstractFunctional dependencies (FDs) play an important role in maintaining data quality. They can be used to enforce data consistency and to guide repairs over a database. In this work, we investigate the problem of missing values and its impact on FD discovery. When using existing FD discovery algorithms, some genuine FDs could not be detected precisely due to missing values or some non-genuine FDs can be discovered even though they are caused by missing values with a certain NULL semantics. We define a notion of genuineness and propose algorithms to compute the genuineness score of a discovered FD. This can be used to identify the genuine FDs among the set of all valid dependencies that hold on the data. We evaluate the quality of our method over various real-world and semi-synthetic datasets with extensive experiments. The results show that our method performs well for relatively large FD sets and is able to accurately capture genuine FDs.

2017

• 
  Harmouch, H., Naumann, F.: Cardinality Estimation: An Experimental Survey.Proceedings of the VLDB Endowment (PVLDB). pp. 499 - 512 (2017).
  [ Abstract ] [ BibTeX ] [ URL ] [ Download ] [ Details ]
   
  Data preparation and data profiling comprise many both basic and complex tasks to analyze a dataset at hand and extract metadata, such as data distributions, key candidates, and functional dependencies. Among the most important types of metadata is the number of distinct values in a column, also known as the zeroth-frequency moment. Cardinality estimation itself has been an active research topic in the past decades due to its many applications. The aim of this paper is to review the literature of cardinality estimation and to present a detailed experimental study of twelve algorithms, scaling far beyond the original experiments. First, we outline and classify approaches to solve the problem of cardinality estimation- we describe their main idea, error guarantees, advantages, and disadvantages. Our experimental survey then compares the performance all twelve cardinality estimation algorithms. We evaluate the algorithms' accuracy, runtime, and memory consumption using synthetic and real-world datasets. Our results show that different algorithms excel in different in categories, and we highlight their trade-offs
  @inproceedings{harmouch2017cardinality, abstract = {Data preparation and data profiling comprise many both basic and complex tasks to analyze a dataset at hand and extract metadata, such as data distributions, key candidates, and functional dependencies. Among the most important types of metadata is the number of distinct values in a column, also known as the zeroth-frequency moment. Cardinality estimation itself has been an active research topic in the past decades due to its many applications. The aim of this paper is to review the literature of cardinality estimation and to present a detailed experimental study of twelve algorithms, scaling far beyond the original experiments. First, we outline and classify approaches to solve the problem of cardinality estimation- we describe their main idea, error guarantees, advantages, and disadvantages. Our experimental survey then compares the performance all twelve cardinality estimation algorithms. We evaluate the algorithms' accuracy, runtime, and memory consumption using synthetic and real-world datasets. Our results show that different algorithms excel in different in categories, and we highlight their trade-offs}, author = {Harmouch, Hazar and Naumann, Felix}, booktitle = {Proceedings of the VLDB Endowment (PVLDB)}, keywords = {hpi profiling distinct_value isg cardinality}, month = 12, number = 4, pages = {499 - 512}, title = {Cardinality Estimation: An Experimental Survey}, volume = 11, year = 2017 }
  Further Information

  URNhttp://dx.doi.org/https://doi.org/10.1145/3164135.3164145

  AbstractData preparation and data profiling comprise many both basic and complex tasks to analyze a dataset at hand and extract metadata, such as data distributions, key candidates, and functional dependencies. Among the most important types of metadata is the number of distinct values in a column, also known as the zeroth-frequency moment. Cardinality estimation itself has been an active research topic in the past decades due to its many applications. The aim of this paper is to review the literature of cardinality estimation and to present a detailed experimental study of twelve algorithms, scaling far beyond the original experiments. First, we outline and classify approaches to solve the problem of cardinality estimation- we describe their main idea, error guarantees, advantages, and disadvantages. Our experimental survey then compares the performance all twelve cardinality estimation algorithms. We evaluate the algorithms' accuracy, runtime, and memory consumption using synthetic and real-world datasets. Our results show that different algorithms excel in different in categories, and we highlight their trade-offs

2016

• 
  Kruse, S., Papenbrock, T., Harmouch, H., Naumann, F.: Data Anamnesis: Admitting Raw Data into an Organization.IEEE Data Engineering Bulletin.39,8-20 (2016).
  [ Abstract ] [ BibTeX ] [ Download ] [ Details ]
   
  Today’s internet offers a plethora of openly available datasets, bearing great potential for novel applications and research. Likewise, rich datasets slumber within organizations. However, all too often those datasets are available only as raw dumps and lack proper documentation or even a schema. Data anamnesis is the first step of any effort to work with such datasets: It determines fundamental properties regarding the datasets’ content, structure, and quality to assess their utility and to put them to use appropriately. Detecting such properties is a key concern of the research area of data profiling, which has developed several viable instruments, such as data type recognition and foreign key discovery. In this article, we perform an anamnesis of the MusicBrainz dataset, an openly available and com- plex discographic database. In particular, we employ data profiling methods to create data summaries and then further analyze those summaries to reverse-engineer the database schema, to understand the data semantics, and to point out tangible schema quality issues. We propose two bottom-up schema quality dimensions, namely conciseness and normality, that measure the fit of the schema with its data, in contrast to a top-down approach that compares a schema with its application requirements.
  @article{kruse2016data, abstract = {Today’s internet offers a plethora of openly available datasets, bearing great potential for novel applications and research. Likewise, rich datasets slumber within organizations. However, all too often those datasets are available only as raw dumps and lack proper documentation or even a schema. Data anamnesis is the first step of any effort to work with such datasets: It determines fundamental properties regarding the datasets’ content, structure, and quality to assess their utility and to put them to use appropriately. Detecting such properties is a key concern of the research area of data profiling, which has developed several viable instruments, such as data type recognition and foreign key discovery. In this article, we perform an anamnesis of the MusicBrainz dataset, an openly available and com- plex discographic database. In particular, we employ data profiling methods to create data summaries and then further analyze those summaries to reverse-engineer the database schema, to understand the data semantics, and to point out tangible schema quality issues. We propose two bottom-up schema quality dimensions, namely conciseness and normality, that measure the fit of the schema with its data, in contrast to a top-down approach that compares a schema with its application requirements.}, author = {Kruse, Sebastian and Papenbrock, Thorsten and Harmouch, Hazar and Naumann, Felix}, journal = {IEEE Data Engineering Bulletin}, keywords = {data_profiling profiling isg data_anamnesis schema_discovery}, month = 6, number = 2, pages = {8-20}, title = {Data Anamnesis: Admitting Raw Data into an Organization}, volume = 39, year = 2016 }
  Further Information

  AbstractToday’s internet offers a plethora of openly available datasets, bearing great potential for novel applications and research. Likewise, rich datasets slumber within organizations. However, all too often those datasets are available only as raw dumps and lack proper documentation or even a schema. Data anamnesis is the first step of any effort to work with such datasets: It determines fundamental properties regarding the datasets’ content, structure, and quality to assess their utility and to put them to use appropriately. Detecting such properties is a key concern of the research area of data profiling, which has developed several viable instruments, such as data type recognition and foreign key discovery. In this article, we perform an anamnesis of the MusicBrainz dataset, an openly available and com- plex discographic database. In particular, we employ data profiling methods to create data summaries and then further analyze those summaries to reverse-engineer the database schema, to understand the data semantics, and to point out tangible schema quality issues. We propose two bottom-up schema quality dimensions, namely conciseness and normality, that measure the fit of the schema with its data, in contrast to a top-down approach that compares a schema with its application requirements.