A little abt me
I have been working on database-related research for close to 5 years.
For many years, data mining has always been an out-of-the-box process, where data is processed out of the DBMS core in order to run the various data mining and maching learning algorithms on it. This does not take advantage of the decades of DBMS research and progress made in query optimisation and processing.
In 2000, researchers at IBM did a detailed study on integrating data mining opeartions into a DBMS to take advantage of the query optimisations and other primitives availanle in a DBMS. At the same time, OLE DB for Data Mining was announced.
At that time, I was working on the kernels for the DBMS called Predator. Then we thot why not? Let's try integrating a common data mining operation into Predator. We came up with Predator-Miner in 2 months, and modified the SQL parser to take in a SQL-variant that allows association rule mining constructs. It was a small prototype which we later demonstrated at one of the international conferences, ICDE 2002.
Fast forward to the present.....
When I first heard the news that data mining will be in Yukon, I was naturally excited! And that begain my journey into dabbling of SQL Server 2005. For the past year, I have been sharing this excitement with fellow database developers using beta-version of the s/w. Finally, in Nov 05, it launched!!!
For many years, data mining has always been an out-of-the-box process, where data is processed out of the DBMS core in order to run the various data mining and maching learning algorithms on it. This does not take advantage of the decades of DBMS research and progress made in query optimisation and processing.
In 2000, researchers at IBM did a detailed study on integrating data mining opeartions into a DBMS to take advantage of the query optimisations and other primitives availanle in a DBMS. At the same time, OLE DB for Data Mining was announced.
At that time, I was working on the kernels for the DBMS called Predator. Then we thot why not? Let's try integrating a common data mining operation into Predator. We came up with Predator-Miner in 2 months, and modified the SQL parser to take in a SQL-variant that allows association rule mining constructs. It was a small prototype which we later demonstrated at one of the international conferences, ICDE 2002.
Fast forward to the present.....
When I first heard the news that data mining will be in Yukon, I was naturally excited! And that begain my journey into dabbling of SQL Server 2005. For the past year, I have been sharing this excitement with fellow database developers using beta-version of the s/w. Finally, in Nov 05, it launched!!!
posted by SQL Server Gems at
9:59 PM
0 Comments:
Post a Comment
<< Home