|
I have a process with 15 effective variables. I could record 9 variables to study its effect on process. Which one is more useful in informing the degree of effectiveness of each variable on scale of 0 to 100? |
|||||
|
It's difficult to tell what is being asked here. This question is ambiguous, vague, incomplete, overly broad, or rhetorical and cannot be reasonably answered in its current form. For help clarifying this question so that it can be reopened, see the FAQ.
|
I assume from the your phrasing "process" that this is time series data. If so with 15 possible supporting variables I guess you want to select the best 9 to use as predictors for some output series. After taking into account any unspecified deterministic structure such as pulses,level shifts,seasonal pulses and/or local time trends you could form a Transfer Function (ARMAX MODEL). After validating that the parameters haven't changed over time and the variance of the errors is constant or at a minimum rendering both of these two assumptions to be true via appropriate Gaussian fixup strategies you might have a useful model. Since you didn't design/ortogonalize the study but simply observed , you can't unambiguosly state how much variability is explained by each input or refer to each input separately as having any partial effect. Collectively the coeffficients ( including any needed lags ) for any and all of the input series can be said to best predict the output series. |
|||
|
|
