|
RapidMiner is available under two different license types:
first, the open-source versions of RapidMiner (Community Edition as well as the Enterprise Edition) which are available under the
GNU Affero General Public License (AGPL)
and a version available under a closed-source (commercial) license for embedding RapidMiner into proprietary products.
Although we try to minimize the number of differences between the open-source and the closed-source versions,
there are some small exceptions which will be discussed here.
In- and Output
| Property |
open-source |
closed-source |
| RapidMiner (YALE) data format |
yes |
yes |
| ARFF |
yes |
yes |
| XRFF |
yes |
yes |
| C4.5 format |
yes |
yes |
| csv |
yes |
yes |
| SVM format (sparse) |
yes |
yes |
| Excel |
yes |
yes |
| SPSS |
yes |
yes |
| Database Input |
yes |
yes |
| dBase |
yes |
yes |
| Text Files |
yes |
yes |
| Audio Files |
yes |
yes |
Machine Learning Algorithms
| Property |
open-source |
closed-source |
| Support Vector Machines (SVM) |
yes |
yes |
| Decision Tree Learners |
yes |
yes |
| Rule Learners |
yes |
yes |
| Lazy Learners |
yes |
yes |
| Bayesian Learners |
yes |
yes |
| Logistic Learners |
yes |
yes |
| Gaussian Processes |
yes |
yes |
| Regression Learners |
yes |
yes |
| Meta Learners |
yes |
yes |
| Association Rule Learning |
yes |
yes |
| Clusterers |
yes |
yes |
Weka Algorithms
Due to complicated license constraints for the Weka toolkit, we decided to replace the most important algorithms
from Weka by own implementations.
The Weka library will remain as part of the open-source version.
The table here indicates which algorithms are replaced by own implementations.
Some of the more exotic algorithms of Weka which are mainly of scientific interest will probably not be replaced.
However, as can be seen in the list above the most important learning schemes will all be available and others are available on request.
| Property |
open-source |
closed-source |
| Learning Schemes |
yes |
all important |
| Attribute Evaluators |
yes |
yes |
| Attribute Weighting Schemes |
yes |
yes |
Data Preprocessing
| Property |
open-source |
closed-source |
| Discretization |
yes |
yes |
| Example Filtering |
yes |
yes |
| Feature Filtering |
yes |
yes |
| Normalization |
yes |
yes |
| Sampling |
yes |
yes |
| Dimensionality Reduction |
yes |
yes |
| Missing (infinite) value handling |
yes |
yes |
| Useless Feature Removal |
yes |
yes |
Feature Operators
| Property |
open-source |
closed-source |
| Feature Selection Schemes |
yes |
yes |
| Feature Weighting Schemes |
yes |
yes |
| Feature Construction Schemes |
yes |
yes |
| Feature Extraction Schemes |
yes |
yes |
| Time Series Handling |
yes |
yes |
Performance Evaluation
| Property |
open-source |
closed-source |
| Cross Validation Schemes |
yes |
yes |
| Leave-One-Out |
yes |
yes |
| Training and Test Sets |
yes |
yes |
| Significance Test |
yes |
yes |
| Set of Performance Criteria |
yes |
yes |
Meta Schemes
| Property |
open-source |
closed-source |
| Parameter Optimization |
yes |
yes |
| Learning Curves |
yes |
yes |
| Process Loops and Iterations |
yes |
yes |
Visualization
| Property |
open-source |
closed-source |
| Online Logging and Plotting |
yes |
yes |
| Large Set of Plotters |
yes |
yes |
| Model Visualization |
yes |
yes |
| High Dimensional Visualizations |
yes |
yes |
| ROC Plots |
yes |
yes |
| Lift Charts |
yes |
yes |
|
