Distributed Data Mining Simulator
 |
Many current data mining tasks can be
accomplished successfully only in a distributed setting. The field of
distributed data mining has therefore gained increasing importance in
the last decade. However, there are still many open questions and
challenges. The Distributed Data Mining Simulator allows to
perform distributed data mining experiments in a simple and flexible
way. The experiments are not actually executed on distributed network
nodes. The tool only simulate this. Simulation makes it easy to
experiment with diverse network structures and communication patterns.
Optimal methods and parameters can be identified efficiently before
putting the system into use. The network structure can for instance be
optimized as part of the general parameter optimization. While this
cannot replace testing the system in an actual network, it makes the
development stage much more efficient.
The distributed data mining simulator is provided as plugin to the open
source data mining suite RapidMiner. |
|
The key features are
- 100% Java implementation
- Simple API
- Very easy to configure and extend
- Combineable with all RapidMiner operators
(e.g. clustering or ensemble learning)
- Supports arbitrary network structures
- Interactive communication inspection and debugging
|
New in Version 1.1:
- Automatic generation of diverse network structures (e.g. scalefree networks, client/server, ...)
- Distributed rule learning
- Distributed feature extraction and sharing
|
Download
The Distributed Data Mining Simulator is currently discontinued. You can access an older version from the Rapid Miner CVS here.
Documentation
The following documentation is
available:
- There is a pdf manual
(pdf), covering the use of the Distributed Data Mining Simulator.
- A good starting point for using the Distributed Data Mining Simulator are the examples provided with the distribution.
Source Code and License
The Distributed Data Mining Simulator is
provided under the
GNU PUBLIC LICENSE (GPL). If you need
another licensing scheme, please contact me. The source oode of the
Distributed Data Mining Simulator can be obtained from the RapidMiner
.
Screenshots
|
The Distributed Data Mining Simulator can be interactively configured through the GUI of RapidMiner and can be combined
with all RapidMiner operators. It is, for instance, possible to cluster an
example set and then to distributed the examples to nodes according to
this clustering.
|
|
For debug purposes, the communication among nodes can be inspected at all points in simulation time.
|
|
The Distributed Data Mining Simulator allows you to create
arbitrary network structures and to visualize them interactively.
|
 |
The Distributed Data Mining Simulator allows you to generate
different network structures and to visualize their properties, as e.g. the distribution of node degrees.
|
Contact
The Distributed Data Mining simulator was created by
Michael Wurst.
Related Links
- For more resources on distributes data mining, refer to the distributed data mining Wiki, maintained by Hillol Kargupta.
- The Distributed Data Mining Simulator is based on the RapidMiner Data Mining framework.
|
|
