(-: (-: SMILES :-):-)

A Stunning Multi-purpose Integrated LEarning System

Presentation:

SMILES is a machine learning system that integrates many different features from other machine learning techniques and paradigms and, more importantly, it presents several innovations in almost all of these features. In particular, SMILES extends classical decision tree learners in many ways (new splitting criteria, non-greedy search, new partitions, extraction of several and different solutions), it has an anytime handling of resources, and has a sophisticated and quite effective handling of (misclassification and test) costs. In this way, SMILES combines and improves the recent interest in hypotheses combination (e.g. boosting) and cost-sensitive learning (a priori and a posteriori class assignments, ROC analysis) outperforming previous systems in many situations. Its applications are basically data-mining and any other machine learning task where decision trees could be useful. It works as a classifier or a probability estimator, especially designed for ranking (ROC analysis and AUC measures). SMILES can also extract comprehensible models from ensembles of classifiers or from neural networks (through mimicking).

The System:

You can download the whole system package for academic use with the following conditions:

DISCLAIMER & COPYRIGHT: The software has been checked on a several Intel-based machines (PCs) under different versions of Linux and Ms. Windows. It 'should' compile on any other system with an ISO C++ compiler possibly under slight modifications. In this regard, you can make any modification to the software, provided you always make the changes explicit and refer to its original authors. Obviously, we are not responsible for any damage caused by the use or misuse of this software. If you find any bug please contact the authors. For commercial use *do* contact the authors.

Source and Executable Code:

Versions for different platforms:

UNIX stable version (Linux): SMILES v.2.3.7 UNIX: Once downloaded and decompressed, follow the readme.txt file and run the shell-script for installation on Unix-like machines (or just run make). It includes the C++ sources. The latest version (under development) is SMILES v.2.6.7 UNIX.
Windows stable version (Console Application for any Win32 operating system): SMILES v.2.3.6 Windows: Program executable compiled with Borland C++ Builder 5.0. Just decompress the zip file and follow the readme.txt file. (Thanks to Ricardo Blanco for this Windows version). The latest version (under development) is SMILES v.2.6.7 Windows.

The system has been successfully compiled in Linux GNU g++ compiler (KDD environment), Borland C++ Builder 5.0 and Microsoft Visual C++ 5.0, so it may be easily portable to any other platform not listed above (in this case download the UNIX version that includes the source files).

MANUAL:

Completely indispensable if you want to take full advantage of SMILES: manual in PDF format.

SAMPLE DATASETS:

Many example datasets in SMILES format can be found here (more than 5Mbs). If you have no examples in SMILES format, please download them because they will be required.

How should it look like?

If the installation is successful, you can directly type:

smiles -?

and you will have the following usage information:

**** SMILES v.2.6.7 (Release Date: 6-September-2003) ****

USAGE:

smiles file.train [file.test] [file.cost] [file.testcost] <file.oracle>

Enjoy SMILES. Let you be surprised by SMILES's learning abilities.

Current Features:

Multitree construction. An AND/OR tree structure is constructed, rather than a forest.
Several Solutions can be selected from the multitree
Solutions can be combined in many different ways.
The extraction of an archetype of the multitree is possible.
Classical features: such as several splitting criteria, pruning, use of expected error and smoothing options (Laplace, m-estimate).
Full validation and evaluation statistics (e.g. cross-validation)
Cost-sensitive and ROC Analysis Features: ROC based and AUC-based splitting criteria and evaluation.
Test cost facilities.
Mimetic methods.
Probability estimation.

Future Features:

Inputs and output in XML format
Incremental extension.
Functional-logic partitions and higher-order extension.