I got a set of data
Time (s) and
Size (Bytes) for each I calculate a
Rate (s/Bytes). I'm trying to predict a Time knowing a Size and using the
Mean and the
3σ of the Rates. Something like
11 minutes 48 seconds ±3σ 2 minutes 30 seconds
For now, to calculate the mean I Sum the Times and the Sizes and calculate the Mean using these two values. But to calculate de Standard Deviation I need to perform the calculation on the calculated Rates of my set of data.
Does that make any sense to calculate the Mean on Times - Sizes sums and the Standard Deviation on each calculated Rates?
Or should I calculate the Mean of each calculated Rates knowing that I'm using a Standard Deviation on each calculated Rates to give ±3σ estimation?
Are both method correct or should I apply the second solution even if the Mean calculation of the first solution is more accurate?