Overview
IQ-Meter is the first comprehensive tool for the evaluation of data-transformation systems. It can facilitate data architects' work on complex scenarios with both commercial and research systems. It is based on three core ideas:
- a definition of a data-transformation system that is sufficiently general to capture a wide class of tools - including research schema-mapping systems, commercial mapping systems and transformation-editors, and ETL tools - and at the same time tight enough for the purpose of our evaluation;
- a new fast algorithm to measure of the quality of the outputs produced by a data translation tool on a given mapping scenario;
- a natural definition and measure of the user-effort needed to achieve such quality.
Ultimately, these techniques allow us to measure the
level of intelligence of a data-transformation tool, defined as the ratio between the
quality of the outputs generated by the system on a given task, and the
amount of user effort required to generate them. The lower the effort required to obtain results of the best quality with a system, the higher its IQ. IQ-Meter builds on these techniques, and provides sophisticated functionalities for the evaluation of data transformation tools.
Download
Last update: 28 Mar 2014
Publications
Related Projects
- BART - Benchmarking Algorithms for (data) Repairing and (data) Translation
- Llunatic - General Framework for Data Cleaning and Schema Mappings
- ++Spicy - Schema Mapping System
People
Comments, Bugs, Requests
