The easiest way to match data from double to integer without losing consistency

I am trying to find the best way to normalize sequentially.

Basically, I have a certain number of instances, each of which has a certain number of attributes with floating values:

For instance:

At1 At2 At3

0.1 0.3 3.0

0.1 4.5 2.1

...

And I want to map each attribute to integer values, trying to be data compatible.

For example, I tried, simply, for each attribute, to separate the difference between the maximum value and the min value for this attribute, dividing it by an arbitrary value of type 10 and then matching all double values ​​of each attribute with the index of its corresponding interval and thereby normalizing my attributes to integer values ​​from 1 to 10 ...

, , , : 1,2, 3,5 223,3 , , , 10 ...

?

+3
1

, encoding compression, , 1-1 .

, , ( , ). .

+2

Source: https://habr.com/ru/post/1779849/


All Articles