• Main Features
  
  
• Structure
  
  
• Graphic Environment
  
  

   KEEL 1.0 is a software tool developed to build and use different Data Mining models. We would like to remark that this is the first software tool of this type containing a library of evolutionary learning algorithms with free code in Java. The main features of KEEL 1.0 are:

• It contains pre-processing algorithms: transformation, discretization, instance selections and feature selections
  
• It also contains a Knowledge Extraction Algorithms library, supervised and unsupervised, remarking the incorporation of multiple evolutionary learning algorithms
  
• It has a statistical analysis library to analyze algorithms
  
• It includes a Java Class library for Evolutionary Computation (JCLEC)
  
• It contains an user-friendly interface, oriented to the analysis of algorithms
  
• The software is used through a web interface, sending to the user all the necessary information to perform the designed experiments in the computer that he wishes

• Experimental Setup
  
• Statistical Analysis Tools
  
• Data Preparation
  
• Knowledge Extraction
  
  

The aim of this part is the design of the desired experimentation over the selected datasets. It provides the following options: Hold Out and Cross Validation.

KEEL 1.0 contains an important library of statistic tests and options for the design of experiments with the goal of analyzing the behaviour of the algorithms.

The pre-processing and knowledge extraction algorithms contained in KEEL are the following:

Pre-processing Knowledge Extraction Algorithms

   The implemented algorithms in KEEL1.0 use the input format of the data file available in the KEEL WEB, and they generate two output files in the output format described in the WEB. The pre-processing algorithms also generate a file with the results of its use, but using the input format.

• Preparation of the Data Bases
  
• Design of Experiments/Execution of Algorithms
  
• Generation of Evolutionary Algorithms with JCLEC library
  

   The Preparation of the Data Bases part makes reference to the previously described tools. It allow the user to create different partitions of his own data bases or the data bases available in the KEEL WEB (KEEL-dataset, described later in this document).

   The Design of Experiments part has the objective of design the desired experiments using a graphical interface. After the experiment is designed, the interface generates a directory structure with all the necessary files needed to run the designed experiments in the local computer selected by the user.

   The interface allows the user to add his own algorithms to the experimentation being designed. The only requirement is to accept the input file format of KEEL. Even, it is not needed to use the Java language for the own algorithms of the user. He only was to put the executable file in the correct directory of the directories structure generated by the interface. This provides a very flexible way for the user to compare his own methods with the ones in KEEL 1.0.

   The Generation of Evolutionary Algorithms with the JCLEC library allows the user to create his own evolutionary algorithms using a graphical interface. The interface uses a language of boxes and a library of implemented methods/routines with the JCLEC library.