Home » excel » excel – Formula used by Weka for Standardization of data

excel – Formula used by Weka for Standardization of data

Posted by: admin May 14, 2020 Leave a comment

Questions:

I am trying to standardize a dataset in Weka. I was able too use to standardize function however the results which I have found using Excel and using Weka were little different.

Following are the screenshots:

Initial Data

enter image description here

Standardized data given by Weka

enter image description here

Manual Excel results

enter image description here

Formula Used for calculating manual results:

z = (x-mean)/SD

z = Std value
x = value to be standardized
mean = mean of dataset
SD = standard deviation

I would like to know the formula which Weka uses to calculate the standardized value.

Any help is appreciated!!
Thanks!!

How to&Answers:

To get the formula that weka usage you need to go through the source code that weka provides.

chekc the souece here @ GC: Standardize – weka.filters.unsupervised.attribute.Standardize (.java) – GrepCode Class Source

You may be able to get the formula you are looking for.

Answer:

I checked using Python and NumPy to do the standardization procedure. Weka (v3.6.12) uses indeed

z = (x - mean) / std

where std is the standard deviation. Be sure that your SD corresponds indeed the standard deviations (and not the variances, for instance).