Quant analytics: nominal scale vs ratio scale
Is there any statistical computation of numeric data or data mining algorithm which varies with scale of the data namely interval scale or ration scale.
For categorical analysis, sometimes algorithm differs from using nominal and ordinal data specially statistical procedure for merging levels of the variable. Is there any statistical procedure of numerical data where the algorithm which varies based on the scale of the data or It just useful for subjective decision making process?
The question is way too general…One quick example I can think of is statistical distances. Mahalonobis distance is for numeric variables and needs some modifications to be generalized on categorical/ordinal variables. And yes, dummy coding is problematic. It is also the key concept in discriminant analysis and clustering algorithm.
I made some mistake on my queries (title/subject headline of my post is miss leading. I am sorry). First of all I want to discuss on comparison between use of interval and ratio scale. I mention nominal and ordinal just as example where I know algorithm differs.
NOTE I now post my TRADING ALERTS into my personal FACEBOOK ACCOUNT and TWITTER. Don't worry as I don't post stupid cat videos or what I eat!