- #1
xeon123
- 90
- 0
I am trying to understand multivariate linear regression.
I have a list of time that it took running processes based on several params, like % of cpu usage, and data read. Eg, I have a process that took 50 seconds to run, with a cpu usage of 70%, and the process read 10bytes of data. I have another process that to 110 seconds to run with a cpu usage of 40% and it read 1Mb of data.
These are ficticious values.With these data, I calculated a multivariate linear regression, and I got 2 coefficients. C1 for the cpu usage, and C2 for the data read.
If i want to predict how much time it will take to run a next process, how the coefficients can help me?
Eg, I know that the CPU is at 30%, and I know that the process will 4Mb of data. Is it correct to predict the time that the process will run like:
t(pred) = C1 * 30% + C2 * 4Mb
Is this correct? I am mixing percentage with Mb to get seconds. Does this make sense? How these params can help me in the prediction?
I have a list of time that it took running processes based on several params, like % of cpu usage, and data read. Eg, I have a process that took 50 seconds to run, with a cpu usage of 70%, and the process read 10bytes of data. I have another process that to 110 seconds to run with a cpu usage of 40% and it read 1Mb of data.
These are ficticious values.With these data, I calculated a multivariate linear regression, and I got 2 coefficients. C1 for the cpu usage, and C2 for the data read.
If i want to predict how much time it will take to run a next process, how the coefficients can help me?
Eg, I know that the CPU is at 30%, and I know that the process will 4Mb of data. Is it correct to predict the time that the process will run like:
t(pred) = C1 * 30% + C2 * 4Mb
Is this correct? I am mixing percentage with Mb to get seconds. Does this make sense? How these params can help me in the prediction?