using rapid miner in browser
Hey guys,
I am currently working on a project for class and I am trying to connect RapidMiner Studio to a MySQL database. Is this possible? In addition, after it is connected, is it possible to control operations done in RapidMiner from a web browser dynamically?
To be more specific, I am using a decision tree to make a prediction on student performance. I want the data provided by student in the database ( this currently collected using a PHP form) to be processed by rapid miner to give the student (on an invidual basis) a predction based on their data and other existing data in the database (the existing data is used to make the model of the decision tree).
Thanks
Answers
Dear Robin,
it is no problem to connect to MySQL. But this requieres the professional edition of RapidMiner. You can get this from our academic program if you would like to use it for your class.
The web browser thing is also possible, but requieres a RM Deployment server. You can then use a POST Webservice to solve your problem.
~Martin
Dortmund, Germany
Actually you can use open source databases like MySQL or PostgreSQL in the Community version.
Im currently using a MySQL database on a WAMP server. I am having problems linking it to the database.
Any idea how to link it?
What's your problem?
You can set up database connections as described e. g. here:
http://docs.rapidminer.com/studio/how-to/mysql-connector.html
Thank you bbarany,
I have gotten to link it.
Now, using the data from the data base (that can now be directly accessed by rapidminer), can i:
1)make an application that runs in the browser (the current one is is php/html)
2)when i click a button creates a decision tree using the data in the data base
3) give a user a prediction what is based on their data using the decision tree made.
My project takes student data and attempts to predict what grade he/she would get in a programming course.
Note: The student would have already entered all of their relavent data in the database.
As mentioned before you can do this using the RM Deployment server as an analytics backend.
Best,
Martin
Dortmund, Germany