SampleClean: Fast and Reliable Analytics on Dirty Data

An important obstacle to accurate data analytics is dirty data in the form of missing, duplicate, incorrect, or inconsistent values. In the SampleClean project, we have developed a new suite of techniques to estimate the results of queries when only a sample of data can be cleaned. Some forms of data corruption, such as duplication, can affect sampling probabilities, and thus, new techniques have to be designed to ensure correctness of the approximate query results. We first describe our initial project on computing statistically bounded estimates of sum, count, and avg queries from samples of cleaned data. We subsequently explored how the same techniques could apply to other problems in database research, namely, materialized view maintenance. To avoid expensive incremental maintenance, we maintain only a sample of rows in a view, and then leverage SampleClean to approximate aggregate query results. Finally, we describe our work on a gradient-descent algorithm that extends the key ideas to the increasingly common Machine Learning-based analytics.

[1]  Léon Bottou,et al.  Stochastic Gradient Descent Tricks , 2012, Neural Networks: Tricks of the Trade.

[2]  Sanjay Krishnan,et al.  Wisteria: Nurturing Scalable Data Cleaning Infrastructure , 2015, Proc. VLDB Endow..

[3]  Tim Kraska,et al.  Stale View Cleaning: Getting Fresh Answers from Stale Materialized Views , 2015, Proc. VLDB Endow..

[4]  Tim Kraska,et al.  Tupleware: Distributed Machine Learning on Small Clusters , 2014, IEEE Data Eng. Bull..

[5]  Sridhar Ramaswamy,et al.  The Aqua approximate query answering system , 1999, SIGMOD '99.

[6]  Joseph M. Hellerstein,et al.  Online aggregation and continuous query support in MapReduce , 2010, SIGMOD Conference.

[7]  F. Olken,et al.  Maintenance of materialized views of sampling queries , 1992, [1992] Eighth International Conference on Data Engineering.

[8]  Jayati.  The Berkeley Data Analytics Stack (BDAS) , 2014, 2014 Conference on IT in Business, Industry and Government (CSIBIG).

[9]  Tim Kraska,et al.  CrowdER: Crowdsourcing Entity Resolution , 2012, Proc. VLDB Endow..

[10]  J. Manthorpe.  Land Registration and Land Valuation in the United Kingdom and in the Countries of the United Nations Economic Commission for Europe (UNECE) , 1998 .

[11]  Ahmed K. Elmagarmid,et al.  Guided data repair , 2011, Proc. VLDB Endow..

[12]  Surajit Chaudhuri,et al.  Optimized stratified sampling for approximate query processing , 2007, TODS.

[13]  Felix Naumann,et al.  The Stratosphere platform for big data analytics , 2014, The VLDB Journal.

[14]  Raymond J. Mooney,et al.  Adaptive duplicate detection using learnable string similarity measures , 2003, KDD '03.

[15]  E. R. Cohen.  An Introduction to Error Analysis: The Study of Uncertainties in Physical Measurements , 1998 .

[16]  Wenfei Fan,et al.  Foundations of Data Quality Management , 2012, Foundations of Data Quality Management.

[17]  Martin L. Kersten,et al.  SciBORQ: Scientific data management with Bounds On Runtime and Quality , 2011, CIDR.

[18]  Helen J. Wang,et al.  Online aggregation , 1997, SIGMOD '97.

[19]  Tim Kraska,et al.  A sample-and-clean framework for fast and accurate query processing on dirty data , 2014, SIGMOD Conference.

[20]  Jennifer Widom,et al.  CrowdFill: collecting structured data from the crowd , 2014, SIGMOD Conference.

[21]  Divesh Srivastava,et al.  Big data integration , 2013, 2013 IEEE 29th International Conference on Data Engineering (ICDE).

[22]  Jun S. Liu,et al.  Metropolized independent sampling with comparisons to rejection sampling and importance sampling , 1996, Stat. Comput..

[23]  Peter Christen,et al.  Febrl: a freely available record linkage system with a graphical user interface , 2008 .

[24]  Ahmed Eldawy,et al.  NADEEF: a commodity data cleaning system , 2013, SIGMOD '13.

[25]  Sanjay Krishnan,et al.  ActiveClean: Interactive Data Cleaning While Learning Convex Loss Models , 2016, ArXiv.

[26]  Frank Olken,et al.  Random Sampling from Databases , 1993 .

[27]  Kun Li,et al.  The MADlib Analytics Library or MAD Skills, the SQL , 2012, Proc. VLDB Endow..

[28]  Theodore Johnson,et al.  Exploratory Data Mining and Data Cleaning , 2003 .

[29]  Doron Rotem,et al.  Simple Random Sampling from Relational Databases , 1986, VLDB.

[30]  Beng Chin Ooi,et al.  Distributed Online Aggregation , 2009, Proc. VLDB Endow..

[31]  Thomas Oberlechner.  Psychology of Judgment and Decision-Making , 2006 .

[32]  Joseph M. Hellerstein,et al.  Quantitative Data Cleaning for Large Databases , 2008 .

[33]  Samuel Madden,et al.  Scorpion: Explaining Away Outliers in Aggregate Queries , 2013, Proc. VLDB Endow..

[34]  Surajit Chaudhuri,et al.  Dynamic sample selection for approximate query processing , 2003, SIGMOD '03.

[35]  Suman Nath,et al.  Tracing data errors with view-conditioned causality , 2011, SIGMOD '11.

[36]  Ion Stoica,et al.  BlinkDB: queries with bounded errors and bounded response times on very large data , 2012, EuroSys '13.

[37]  E. H. Simpson,et al.  The Interpretation of Interaction in Contingency Tables , 1951 .

[38]  Alon Y. Halevy,et al.  Pay-as-you-go user feedback for dataspace systems , 2008, SIGMOD Conference.

[39]  Chris Jermaine,et al.  Online aggregation for large MapReduce jobs , 2011, Proc. VLDB Endow..

[40]  Paolo Papotti,et al.  Descriptive and prescriptive data cleaning , 2014, SIGMOD Conference.

[41]  Jianzhong Li,et al.  Towards certain fixes with editing rules and master data , 2010, The VLDB Journal.

[42]  Minos N. Garofalakis,et al.  Approximate Query Processing: Taming the TeraBytes , 2001, VLDB.

[43]  Jeffrey F. Naughton,et al.  Corleone: hands-off crowdsourcing for entity matching , 2014, SIGMOD Conference.

[44]  Paolo Papotti,et al.  KATARA: A Data Cleaning System Powered by Knowledge Bases and Crowdsourcing , 2015, SIGMOD Conference.