Mapping and Cleaning
by Prof. Gianni Mecca, University of Basilicata
April 28th 2015 @ 9:15, in Room Levico
In this invited talk Prof. Gianni Mecca will present his work about
Mapping and Cleaning
Data-cleaning (or data-repairing) is considered a crucial problem in many database-related tasks. It consists in making a database consistent with respect to a set of given constraints. Most existing methods repair data for specific classes of constraints only, and rely on ad-hoc decisions to heuristically find a single repair. In the seminar, we present Llunatic, a uniform framework that generalizes existing approaches by seeing repairs as solutions of so-called cleaning scenarios. Furthermore, we develop a chase-based algorithm to compute multiple solutions, and show how existing heuristics can be incorporated into the chase procedure in a principled way. A number of techniques are proposed that improve the scalability of the chase algorithm. Based on these, we have implemented the framework in a DBMS-based prototype, and we report experimental results that confirm its good scalability and superior quality in computing repairs. We conclude the paper by showing that the framework can be extended to handle mapping and cleaning tasks in an integrated fashion.
Giansalvatore Mecca is full professor at the Faculty
of Sciences of Università della Basilicata, where he currently serves also as Deputy Rector for Research.
Giansalvatore Mecca, born in Potenza, is full professor at the Faculty of Sciences of Università della Basilicata, where he currently serves also as Deputy Rector for Research. He graduated with honors in Computer Engineering from Università di Roma "La Sapienza" and received his PhD, also from "La Sapienza", under the supervision of Paolo Atzeni. He is with Universita' della Basilicata from 1995, first as a research associate and then as an associate professor. He also had visiting appointments at the University of Toronto and Universita' di Roma Tre. His research interests include information extraction and data management techniques for XML and Web data. He has also worked on cooperative database systems, string databases, deductive databases, and object-oriented databases.
He teaches various courses in Computer Programming, Databases, and Web Development, for the Computer Science diploma of Università della Basilicata.
Contact Info: Matteo Lissandrini (firstname.lastname@example.org)