Several question on Rapidminer

jesslyn

hi. i am currently evaluating some data mining software and i want to know more about rapidminer.

1) how rapidminer provide better multiple interfaces?
2) how good can rapidminer in accessing and managing the data? (how well it can reason with noise level)
3) does rapidminer covers all data mining algorithms? which are the algorithms RM is covering?

IngoRM

Dear Jesslyn,

I will try to answer your questions:

1) how rapidminer provide better multiple interfaces?

If you refer to interfaces in the sense of interacting with RapidMiner: yes, there are several options. These include executing analytics processes in the graphical user interface of your desktop, executing processes in command line / batch mode, executing processes remotely on the RapidAnalytics server, transforming processes into web services which can then be integrated into your own infrastructure, and finally you can interact with RapidMiner via a Java API.

If you refer to data source interfaces, RapidMiner is probably one of the products in the market with most connectors to data sources. RapidMiner can connect to dozens of file formats, databases, and web services. More information can be found in the fact sheet (link is below).

2) how good can rapidminer in accessing and managing the data? (how well it can reason with noise level)

RapidMiner provices far more than 1000 basic operations (which we call "operators") plus several extensionjs. About 80% of all of these operators of RapidMiner are about data preprocessing, many of them also for noise handling. More information can be found in the fact sheet (link below).

3) does rapidminer covers all data mining algorithms? which are the algorithms RM is covering?

Basically: yes. RapidMiner comes with a lot of data mining algorithms already and there are extensions which cover specific topics like frequent itemset mining even better. Then there is an extension which integrates all algorithms from Weka and another one for connecting with, also coming with hundreds of algorithms. Hence, RapidMiner definitely cover much more algorithms than any other system, especially other commercial systems which usually only provide a small selection of data mining techiques. Again: More can be found in the fact sheet below 


Fact Sheet

Probably, you will find the following fact sheet for RapidMiner and RapidAnalytics interesting:

    http://rapid-i.com/downloads/rapidminer/facts/rapidminer_rapidanalytics_fact_sheet.pdf

It covers many of the details and functions of RapidMiner and points also out some unique features. Hope that helps. Cheers,
Ingo

jesslyn

I have some question.
Does the community edition covers all the approaches and technique in the Enterprise Standard version? Or limited only?

IngoRM

Hi again,

no, there are differences between the Community Edition and the Enterprise Editions. For the Enterprise Editions, we offer some additional features:

• In-Database Mining: Offers several modeling schemes which are directly trained in database without loading the data
• Profiler: Showing the amount of time used in different operators, aggregations over runtimes, and a memory monitor
• Data Editor: Excel-like data editor for repository entries, allows for data changes and meta data corrections
• Access to OLAP cubes
• Access to SAP
• amongh others

The next versions of the Enterprise Editions will also offer an improved documentation mechanism including post-its directly in the process editor, a new extension for real stream mining, and a new mechanism for creating new operators based on subprocesses (operator libraries).

And we of course offer guarantees and support to our Enterprise customers. More information can be found here:

    http://rapid-i.com/content/view/181/190/


The differences for RapidAnalytics are even larger. RapidAnalytic Enterprise Edition offer web-based report designers, dashboards, brand styling, style bundles among others. More information can be found at

http://rapid-i.com/content/view/182/192/

as well as in our introduction videos on our web site.

Cheers,
Ingo

jesslyn

I mean the approaches(neural network, decision tree) and techniques(classification, regression). does the community edition have all of it? or only limited approaches and techniques?

IngoRM

Hi,

the learning methods available in the Community Edition are not limited (in the sense that they are performing worse), but there are simply even more operators in the Enterprise Edition. For example those which are directly performing a decision tree learning or Naive Bayes in the database. And there are more connectors to different data sources. The other additional features are more in the areas of user support and professional analysis in collaborative environments like improved documentation etc.

Cheers,
Ingo