Download A High-Performance Data Mining Framework in MySQL

Survey
yes no Was this document useful for you?
   Thank you for your participation!

* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project

Document related concepts

Cluster analysis wikipedia , lookup

Nonlinear dimensionality reduction wikipedia , lookup

Transcript
1 of 7
A High-Performance Data
Mining Framework in MySQL
Dr. Lutz Hamel
Tiegeng Ren
Dept. of Computer Science in URI
[email protected]
3/31/2003
2 of 7
A High-Performance Data
Mining Framework in MySQL
•
•
•
•
The idea
What we have
Bring it to MySQL framework
Current status
3 of 7
The Idea
•
•
•
•
High performance database
Mature data mining algorithms
Current solutions – use DB interface.
Can we bypass the interface/data transfer?
4 of 7
What We Have
• High-performance relational database
• Data mining algorithms
Database
Data mining
Algorithms (ANN,
Decision Tree …)
5 of 7
Comparison of Solutions
Database (Oracle, Interface (ODBC, User functions/Algorithm
MSSQL, ySQL..) JDBC, APIs)
(Data mining algorithms)
Data transfer in/out DB through slow
interface.
Data mining Algorithms
(ANN, Decision Tree …)
Database (Oracle,
MSSQL, ySQL..)
• Data managed by the High-performance
DB itself.
• No data transfer any more.
• Algorithms are wrapped into functions
that available to all SQL queries.
6 of 7
How to integrate data mining
algorithm to RDBMS
• Why MySQL
• Implementation
– Interface - User Defined Functions (UDF)
– ANN algorithms
• What we plan to do
– Implant a prototype ANN into MySQL as a UDF
– Enhance the interface
– Put in more algorithms and functionality
7 of 7
Current Status
• Probe the MySQL
• Redesign the interface of current ANN
program into the UDF
– Add UDF wrapper function to ANN.
– Modify the ANN’s interface