|
Category Mapping
GeneXproTools supports all kinds of categorical variables, both as part of entirely
categorical datasets or intermixed with numerical variables. In both cases the
categories in all categorical variables are automatically replaced by numerical values
so that you can start modeling straightaway.
GeneXproTools uses simple heuristics to make this initial mapping, but then lets you choose
more meaningful mappings in the Category Mapping Window.
Dependent categorical variables are also supported but they are handled differently in
classification and logistic regression problems with more than two classes.
In these cases the
mapping is made so that only one class is singled out, resulting in a binomial outcome,
such as {0, 1} or {-1, 1}, which can then be used to create classification or
logistic regression models. The merging of the response variable in classification and
logistic regression is handled in the Class Merging & Discretization Window.
In regression problems, dependent categorical variables are handled exactly as any other
categorical variable, that is, the categories in the response variable are also converted
to numerical values using user-defined mappings.
The beauty and power of GeneXproTools support for categorical variables goes beyond
giving you access to a sophisticated and extremely useful tool for changing and
experimenting with different mappings easily and quickly by trying out different scenarios
and seeing immediately how they impact on modeling. Indeed GeneXproTools also generates code
that supports data in exactly the same format that was loaded into GeneXproTools. This means
that all the code generated both for external model deployment or for scoring internally in
GeneXproTools, also supports categorical variables. Below is an example in C++ of a
logistic regression model with 12 variables, 7 of which are categorical.
//------------------------------------------------------------------
// Logistic regression model generated by GeneXproTools 5.0 on 5/17/2013 4:13:11 PM
// GEP File: D:\GeneXproTools\Version5.0\OnlineGuide\LoanRisk_03a.gep
// Training Records: 667
// Validation Records: 333
// Fitness Function: Bounded ROC, Logistic Threshold
// Training Fitness: 726.124343110239
// Training Accuracy: 76.76% (512)
// Validation Fitness: 777.537892479522
// Validation Accuracy: 79.58% (265)
//------------------------------------------------------------------
#include "math.h"
#include "string.h"
double gepModel(char* d_string[]);
double gep3Rt(double x);
void TransformCategoricalInputs(char* input[], double output[]);
double gepModel(char* d_string[])
{
const double G3C5 = -4.97848445081942;
double d[20];
TransformCategoricalInputs(d_string, d);
double dblTemp = 0.0;
dblTemp = (pow(d[0],4)+d[15]);
dblTemp += exp(d[2]);
dblTemp += (((d[1]*pow(gep3Rt((d[5]+G3C5)),3))-d[18])-d[8]);
dblTemp += ((d[9]*(d[10]+d[12]))-(((d[7]*d[11])*d[7])*gep3Rt(d[10])));
const double SLOPE = 6.9596006314631E-03;
const double INTERCEPT = 3.45748287482188E-02;
double probabilityOne = 1.0 / (1.0 + exp(-(SLOPE * dblTemp + INTERCEPT)));
return probabilityOne;
}
double gep3Rt(double x)
{
return x < 0.0 ? -pow(-x,(1.0/3.0)) : pow(x,(1.0/3.0));
}
void TransformCategoricalInputs(char* input[], double output[])
{
if(strcmp("A11", input[0]) == 0)
output[0] = 1.0;
else if(strcmp("A12", input[0]) == 0)
output[0] = 2.0;
else if(strcmp("A13", input[0]) == 0)
output[0] = 3.0;
else if(strcmp("A14", input[0]) == 0)
output[0] = 4.0;
else output[0] = 0.0;
output[1] = atof(input[1]);
if(strcmp("A30", input[2]) == 0)
output[2] = 1.0;
else if(strcmp("A31", input[2]) == 0)
output[2] = 2.0;
else if(strcmp("A32", input[2]) == 0)
output[2] = 3.0;
else if(strcmp("A33", input[2]) == 0)
output[2] = 4.0;
else if(strcmp("A34", input[2]) == 0)
output[2] = 5.0;
else output[2] = 0.0;
if(strcmp("A61", input[5]) == 0)
output[5] = 1.0;
else if(strcmp("A62", input[5]) == 0)
output[5] = 2.0;
else if(strcmp("A63", input[5]) == 0)
output[5] = 3.0;
else if(strcmp("A64", input[5]) == 0)
output[5] = 4.0;
else if(strcmp("A65", input[5]) == 0)
output[5] = 5.0;
else output[5] = 0.0;
output[7] = atof(input[7]);
if(strcmp("A91", input[8]) == 0)
output[8] = 1.0;
else if(strcmp("A92", input[8]) == 0)
output[8] = 2.0;
else if(strcmp("A93", input[8]) == 0)
output[8] = 3.0;
else if(strcmp("A94", input[8]) == 0)
output[8] = 4.0;
else output[8] = 0.0;
if(strcmp("A101", input[9]) == 0)
output[9] = 1.0;
else if(strcmp("A102", input[9]) == 0)
output[9] = 2.0;
else if(strcmp("A103", input[9]) == 0)
output[9] = 3.0;
else output[9] = 0.0;
output[10] = atof(input[10]);
if(strcmp("A121", input[11]) == 0)
output[11] = 1.0;
else if(strcmp("A122", input[11]) == 0)
output[11] = 2.0;
else if(strcmp("A123", input[11]) == 0)
output[11] = 3.0;
else if(strcmp("A124", input[11]) == 0)
output[11] = 4.0;
else output[11] = 0.0;
output[12] = atof(input[12]);
output[15] = atof(input[15]);
if(strcmp("A191", input[18]) == 0)
output[18] = 1.0;
else if(strcmp("A192", input[18]) == 0)
output[18] = 2.0;
else output[18] = 0.0;
}
See Also:
Related Tutorials:
Related Videos:
|
