Changing numbers in dataset

Hi,

I have a dataset of size m*n. The dataset only contains integer numbers between 1 to 5. Now I want to change the numbers as

1=-1

2=-1+(2/4)=-1/2

3=-1+(4/4)=0

4=-1+(64)=1/2

5=-1+(8/4)=1

Is there any efficient way of doing that?

Answer 1

[Voss]

Considering that -1 can be written as -1+(0/4), a pattern becomes clear:

1 -> -1+(0/4)

2 -> -1+(2/4)

3 -> -1+(4/4)

4 -> -1+(6/4)

5 -> -1+(8/4)

In general, say:

x -> -1+(y/4)

The pattern is that each time x is incremented by 1 on the left-hand side: x = (1, 2, 3, 4, 5), that corresponds to y being incremented by 2 on the right-hand side: y = (0, 2, 4, 6, 8). And y is 0 when x is 1.

Thus we can say that y = 2*(x-1), so that the rule is:

x -> -1+(2*(x-1)/4)

or, simplifying:

x -> -1+(x-1)/2

So if x is your m-by-n matrix, then the transformed version of x is -1+(x-1)/2.

m = 3;
n = 4;
x = randi(5,[m,n])

x = 3×4

4 2 5 2 3 2 1 2 1 5 5 2

new_x = -1+(x-1)/2

new_x = 3×4

0.5000 -0.5000 1.0000 -0.5000 0 -0.5000 -1.0000 -0.5000 -1.0000 1.0000 1.0000 -0.5000

[Rajesh]

Thank you very much!

[Voss]

You're welcome!

Answer 2

[Chris]

For only five values, I think it's pretty efficient to do a direct replacement:

ds_new = zeros(size(dataset))
ds_new(dataset == 1) = -1;
ds_new(dataset == 2) = -0.5;
% ...etc

If you need more flexibility, this is what I came up with:

ds = randi(5,4)

ds = 4×4

4 4 1 2 4 2 5 4 3 1 2 5 1 3 2 3

ints = 1:5;
vals = 0.5*(ints-1)-1;

dsnew = zeros(size(ds));
for idx = 1:numel(ints)
    dsnew(ds == idx) = vals(idx);
end
dsnew

dsnew = 4×4

0.5000 0.5000 -1.0000 -0.5000 0.5000 -0.5000 1.0000 0.5000 0 -1.0000 -0.5000 1.0000 -1.0000 0 -0.5000 0