TIMi .NET Library v1.02
=======================

This ZIP archive contains a C#/.NET library that allows using TIMi predictive
models directly inside a .NET application. The library provides a simple class
that is named "TIMiModel".

Here is the declaration of the "TIMiModel" class:
  public class TIMiModel
  {
      TIMiModel(String^ filename, array<String^> ^columnNames);
      double eval(array<String^> ^oneRow);
  }

The objective of this class is to apply a TIMi model on a data-table, row 
by row. Each call to the "eval" method receives one row of the data-table 
and returns a prediction.

The constructor uses 2 parameters:
1. The name of the .ModelXML file containing the TIMi model.
2. An array containing the column names of the data-table.
   Based on the column names, the library will "select" only the relevant 
   columns when computing a new prediction.

The only input parameter of the "eval" method is an array of strings that
represents one row of the data-table. The order of the columns in this 
array must match exactly the order given in the array-parameter 
"columnNames" given when calling the constructor of the object.

The directory:
 * "debugWin32"   contains the debug  -32 bit version of the library.
 * "releaseWin32" contains the release-32 bit version of the library.
 * "debugX64"     contains the debug  -64 bit version of the library.
 * "releaseX64"   contains the release-64 bit version of the library.

The directory "exampleUsage" contains a very small .NET managed project
that illustrates how to use to library.

NOTE
====
The decimal separator that is used inside the fields containing floating 
point numbers is always the dot (".")(This is independent of the MSWindows 
settings).
