Finding the Mode in a Vector of Floats - c++

Question

I am trying to find the mode average in a vector containing 324 float values.
The code I have is as follows:
float max = vec.back();
float prev = max;
float mode = 0.0;
int maxcount = 0;
int currcount = 0;
for (const auto n : vec) {
if (n == prev) {
++currcount;
if (currcount > maxcount) {
maxcount = currcount;
mode = n;
}
} else {
currcount = 1;
}
prev = n;
}
std::cout << mode << std::endl
This prints out the mode to be 0.75, which is wrong.
Here are all the float values, they come from a txt file so please excuse the format:
0.61 0.61 0.61 0.62 0.62 0.62 0.62 0.62 0.62 0.62 0.63 0.63 0.63 0.63 0.63 0.63 0.63 0.63 0.63 0.63 0.63 0.63 0.63 0.63 0.63 0.63 0.63 0.63 0.63 0.63 0.63 0.63 0.63 0.63 0.63 0.63 0.63 0.63 0.63 0.63 0.63 0.63 0.63 0.64 0.64 0.64 0.64 0.64 0.64 0.64 0.64 0.64 0.64 0.64 0.64 0.64 0.64 0.64 0.64 0.64 0.64 0.64 0.64 0.64 0.64 0.65 0.65 0.65 0.65 0.65 0.65 0.65 0.65 0.65 0.65 0.65 0.65 0.65 0.65 0.65 0.65 0.65 0.65 0.65 0.65 0.65 0.65 0.65 0.65 0.65 0.65 0.65 0.65 0.65 0.65 0.65 0.65 0.65 0.65 0.65 0.65 0.65 0.65 0.65 0.65 0.65 0.66 0.66 0.66 0.66 0.66 0.66 0.66 0.66 0.66 0.66 0.66 0.66 0.66 0.66 0.66 0.66 0.66 0.66 0.66 0.67 0.67 0.67 0.67 0.67 0.67 0.67 0.67 0.67 0.67 0.67 0.67 0.67 0.67 0.67 0.67 0.67 0.67 0.67 0.67 0.68 0.68 0.68 0.68 0.68 0.68 0.68 0.68 0.68 0.68 0.68 0.68 0.68 0.68 0.68 0.68 0.68 0.68 0.68 0.68 0.68 0.68 0.68 0.68 0.68 0.68 0.68 0.68 0.68 0.68 0.68 0.68 0.68 0.68 0.68 0.68 0.69 0.69 0.69 0.69 0.69 0.69 0.69 0.69 0.69 0.69 0.69 0.69 0.69 0.69 0.69 0.69 0.69 0.69 0.69 0.69 0.69 0.7 0.7 0.7 0.7 0.7 0.7 0.7 0.7 0.7 0.7 0.7 0.7 0.7 0.7 0.7 0.7 0.7 0.7 0.7 0.7 0.71 0.71 0.71 0.71 0.71 0.71 0.71 0.71 0.71 0.71 0.71 0.71 0.71 0.71 0.72 0.72 0.72 0.72 0.72 0.72 0.72 0.72 0.72 0.72 0.72 0.72 0.72 0.72 0.72 0.72 0.72 0.72 0.72 0.72 0.73 0.73 0.73 0.73 0.73 0.73 0.73 0.73 0.73 0.73 0.73 0.73 0.73 0.74 0.74 0.74 0.74 0.74 0.74 0.74 0.75 0.75 0.75 0.75 0.75 0.75 0.75 0.75 0.75 0.75 0.75 0.76 0.76 0.76 0.76 0.76 0.76 0.76 0.76 0.76 0.77 0.77 0.77 0.77 0.77 0.77 0.77 0.78 0.78 0.78 0.78 0.78 0.78 0.78 0.78 0.78 0.78 0.79 0.79 0.79 0.79 0.79 0.79 0.79 0.79 0.79 0.79 0.79
Excel presents the mode as 0.65. Why does my code not produce the same result? What do I need to change?
Many thanks.
edit: I have found through debugging the values within vec are more like; 0.68000000000000005, 0.69999999999999996, though some are still only two decimal points (0.64, 0.74 etc). Could this be the issue? Am I able to round up the values for this particular calc?

Answer 1

The problem might be the use of floats for comparison. Because of how they are stored, floating point numbers differ, in general, from the value they are initialized to by a small amount.
Instead of using n == prev, consider a comparison within some small epsilon that is greater than the machine precision (for any machine you expect to run this code on) but less than the smallest true difference between any of your two numbers (which looks like 0.01). So you could do
if (((n - prev) < EPSILON) && ((prev - n) < EPSILON)) { ...`
for float EPSILON = 0.000001, or a value that makes sense for you.
See also this question on comparing floats. Of note is that the ideal epsilon would change if your data set changed to much larger or much smaller numbers.
Even if there is another problem in your code, you might consider moving away from comparing floats in general.

Answer 2

By debugging I found that my values were not just two decimal place values, therefore, the mean average was actually 0.7500000000004, but was still being printed as 0.75.
By adding a rounding function call, and removing the const I was able to find the mean to two decimal places.
for (auto n : vec)
{
n = roundf(n * 100) / 100;
if (n == prev)
{
++currcount;
if (currcount > maxcount)
{
maxcount = currcount;
mode = n;
}
} else
{
currcount = 1;
}
prev = n;
}