The output numpy array of my program looks like this:
[0.8500000000000001, 0.9000000000000001, 0.8800000000000001, 0.875]
However, I want to get them as;
0.85,0.9,0.88,0.875
Any idea, why that happens and how to get rid of it?
Asked
Active
Viewed 19 times
0
-
1floating-point arithmetic is imprecise. – juanpa.arrivillaga Sep 21 '17 at 02:24
-
How can I get rid of it? – Sep 21 '17 at 02:25
-
1You can't if you are using floating-point arithmetic. – juanpa.arrivillaga Sep 21 '17 at 02:25
-
@juanpa.arrivillaga You can if you are using *decimal* floating point arithmetic. – Stefan Pochmann Sep 21 '17 at 02:31
-
@StefanPochmann Well, you can get rid of this particular example of imprecision, but imprecision remains with decimal floating point arithmetic, no? – juanpa.arrivillaga Sep 21 '17 at 02:34
-
@juanpa.arrivillaga In general, yes, like when you compute 1/3. But those numbers look to me like they were created in a way such that using decimals would've prevented the problem. – Stefan Pochmann Sep 21 '17 at 02:42