Welcome to ENTOOL

Introduction

ENTOOL is a software package for ensemble regression and classification.

It is implemented mainly in Matlab, with some time-critical parts written in C/C++ (as mex-functions).

CTOOL is a fork of entool for classification, now available in Octave

Objectives

• Extending the ensemble learning approach to several types of models (hybrid ensembles)
• Object-oriented implementation yields a transparent mixture of different models and allows the user the addition of his own model classes

Methods

The toolbox is equipped with several model classes for out-of-box usage:

• Radial basis functions (RBF)
• Linear regression (LIBLINEAR)
• Polynomial regression
• K-nearest-neighbour models with adaptive metric (VICINAL)
• Multilayer perceptron (MLP)
• Adaptive Regression Splines (ARES)
• Classification and regression trees (CART)
• Linear Discriminant Analysis (LDA)
• Penalized Discriminant Analysis (PDA)
• Support Vector Machines (SVM)
• Logistic regression (LOGREG)
• Ridge regression (KRIDGE)
• Gaussian Mixture Models (GMM)

What do I need to run ENTOOL ?

1. Matlab version 6.0 or newer
2. An unpacking utility like Winzip or gunzip and tar
3. If mex-files for your platform are not included, an ANSI C++ compliant compiler (e.g. gcc 3.4) is needed to compile the mex-files from the source code.

How to install ENTOOL on my computer ?

• ENTOOL is based on Matlab, so it will run on any platform for which Matlab is available. Some parts of ENTOOL are coded in C++ and require therefore an ANSI C++ compliant compiler (e.g. gcc 3.4) to be compiled. Precompiled binaries are included for Windows (95/98/NT) and Linux.
• Download and install CTOOL for classification (Octave version)
• Download and install ENTOOL for regression Version 1.1 (This version is not longer supported)
• Download and install ENTOOL for regression (Version 1.0)

References and use case scenarios

This is the toolbox that we used successfully for several tasks and in different challenges:

Environmental Toxicity Prediction Challenge
One of the first past winners

J. Wichard, R. Kuehne
Predicting aqueous solubility from structure
Journal of the University of Applied Sciences Mittweida,
Proceedings of the 20. IWKM, 28.-29. Oct 2009.

J.D. Wichard
Classification of Ford Motor Data
Contribution to the Ford Classification Challenge at WCCI 2008, Hong Kong.

J.D. Wichard, H. Cammann, C. Stephan and T. Tolxdorff
Classification Models for Early Detection of Prostate Cancer
Journal of Biomedicine and Biotechnology, Volume 2008, Article ID 218097, doi:10.1155/2008/218097.

J.D. Wichard
Model Selection in an Ensemble Framework
Proceedings of the IEEE World Congress on Computational Intelligence WCCI 2006, pp. 2187- 2192, Vancouver, Canada.

Andreas Rothfuss, Thomas Steger-Hartmann, Nikolaus Heinrich, and Jörg Wichard
Computational Prediction of the Chromosome-Damaging Potential of Chemicals
Chemical Research in Toxicology, 19 (10), pp. 1313 -1319, September 2006.

J.D. Wichard
Agnostic Learning with Ensembles of Classifiers
Proceedings of the IJCNN 2007, Orlando, Florida, USA.

Contact

Or simply Email: joerg.wichard (at) gmail.com

If you have compiled the mex-files for a platform that is not supported in this distribution, please send us those so that we can make them available for public use.

Copyright notice

ENTOOL now falls under the Gnu Public License (GPL).

Acknowledgements

The work was done within the Research Training Network COSYC of SENS No. HPRN-CT-2000-00158 within in 5th EU Framework Program of the European Community.

---

Copyright © 2002-2012 J.D. Wichard and C. Merkwirth