In week 1: Modelling complex data

In modelling complex data, where we give x as input; x being[age,weight,ht, bloodpressur…]
and we put it in a non linear function f(x), Here says an statement, "I am not interested in answers how much blood sugar level depend on the weight or age?
My question is It does seem plausible, Can there be an argument where we need like what’s the no of overweight people who got blood pressure so we infer obese people have more likely to have blood sugar.

I would like to answer using this quote
"Torture the data, and will confess to anything "
The first and for most important task in data science is “Question that you wanna ask from data set”.To know whether there is a relation between overweight and blood sugar.You require domain knowledge related to that data set. If there is any proved relation between overweight , blood pressure and blood sugar. Obviously you can infer obese people have more likely to have blood sugar.

This is a similar question I asked before. The reply goes like this:-
The relation between the parameters can be found(if there is any) using Exploratory Data Analysis. There are some simple relationships which can find using statistical analysis which is more transparent and easy to understand, however there are some complex relationships(complex functions) where we require algorithmic analysis where the role of machine learning comes, which is again data hungry… Therefore it depends what you want from data…

1 Like