calculating gini index

2015228094

I want to what type of Gini coeffcient used in item distribution and what is the formula to calculate Gini coeffcient and Gini index

MartinLiebig

Hi,

 

the one used in Trees and Weight By Gini can be found here:

 

https://github.com/rapidminer/rapidminer-studio/blob/master/src/main/java/com/rapidminer/operator/learner/tree/criterions/GiniIndexCriterion.java

 

Best,

Martin

2015228094

Thank you

Is this formula used for clustring kmean evalution?

do you have the formula as mathematical equation?

MartinLiebig

Hi,

 

for clustering its: https://github.com/rapidminer/rapidminer-studio/blob/master/src/main/java/com/rapidminer/operator/validation/clustering/exampledistribution/GiniCoefficient.java

or in code

		double sum = 0;
		for (int i = 0; i < x.length; i++) {
			sum = sum + x[i];
		}

		double mean = sum / n;

		sum = 0;
		for (int i = 0; i < x.length; i++) {
			for (int j = 0; j < x.length; j++) {
				sum = sum + Math.abs(x[i] - x[j]);
			}
		}

		if (mean == 0) {
			return 0.0;
		}

		return 1 - (sum / (2 * mean * (n * ((double) n - 1))));

 

Best,

Martin

 

MartinLiebig

Hi,

 

the one used in Trees and Weight By Gini can be found here:

 

https://github.com/rapidminer/rapidminer-studio/blob/master/src/main/java/com/rapidminer/operator/learner/tree/criterions/GiniIndexCriterion.java

 

Best,

Martin

2015228094

Thank you

Is this formula used for clustring kmean evalution?

do you have the formula as mathematical equation?

MartinLiebig

Hi,

 

for clustering its: https://github.com/rapidminer/rapidminer-studio/blob/master/src/main/java/com/rapidminer/operator/validation/clustering/exampledistribution/GiniCoefficient.java

or in code

		double sum = 0;
		for (int i = 0; i < x.length; i++) {
			sum = sum + x[i];
		}

		double mean = sum / n;

		sum = 0;
		for (int i = 0; i < x.length; i++) {
			for (int j = 0; j < x.length; j++) {
				sum = sum + Math.abs(x[i] - x[j]);
			}
		}

		if (mean == 0) {
			return 0.0;
		}

		return 1 - (sum / (2 * mean * (n * ((double) n - 1))));

 

Best,

Martin