intersect with a confidence interval?

Hi,

Let's assume I have array a=[1.2345,1.4,1.2] and b=[1.2344,1.4,1.3]...

is there any quick way to define that if the array members have a difference less than 0.0001 they should appear in the intersection? (in this example intersect(a,b)=[1.2345,1.4])

thanks

Answer 1

[Star Strider]

The intersect function doesn’t allow tolerances.

You can ‘sort of’ get around that by making your own function to truncate the numbers to a specific precision, then use intersect:

a=[1.2345,1.4,1.2];
b=[1.2344,1.4,1.3];

floorn = @(x,n) floor(x.*10^n)/10^n;

c = intersect(floorn(a,3), floorn(b,3))

If you choose to have intersect return the indices, you can then recover the original numbers from the intersection:

[c,ia,ib] = intersect(floorn(a,3), floorn(b,3));
result = a(ia)

produces:

result =

         1.2345          1.4

although b(ib) would return the corresponding values in ‘b’, [1.2344 1.4]. That will be your choice.