The Alternating Decision Tree Learning Algorithm
ICML '99 Proceedings of the Sixteenth International Conference on Machine Learning
Generic Schema Matching with Cupid
Proceedings of the 27th International Conference on Very Large Data Bases
Model management 2.0: manipulating richer mappings
Proceedings of the 2007 ACM SIGMOD international conference on Management of data
Summarizing relational databases
Proceedings of the VLDB Endowment
Data integration for the relational web
Proceedings of the VLDB Endowment
Proceedings of the VLDB Endowment
Hi-index | 0.00 |
Motivated by eScience applications, we explore automatic generation of example "starter" queries over unstructured collections of tables without relying on a schema, a query log, or prior input from users. Such example queries are demonstrably sufficient to have non-experts self-train and become productive using SQL, helping to increase the uptake of database technology among scientists. Our method is to learn a model for each relational operator based on example queries from public databases, then assemble queries syntactically operator-by-operator. For example, the likelihood that a pair of attributes will be used as a join condition in an example query depends on the cardinality of their intersection, among other features. Our demonstration illustrates that datasets with different statistical properties lead to different sets of example queries with different properties.