Reply With QuoteTo do what you're suggesting requires a lot of questionable assumptions if you only have 2 data points ! If you have many units of furniture, it's better and you could do a linear or nonlinear regression.
Their furniture units have a number of wooden panels & pieces NP and a number of fasteners NF.
If I assemble two units of furniture in times X and Y I have two equations and two unknowns and I can find the relationship between NP, NF and assembly time and so calculate assembly time based only on NP or on NF or on both.
With many units of furniture, how should I average the values of NP and NF?
How would I find meaningful standard deviations for assembly time based the NP and NF values?
To do what you're suggesting requires a lot of questionable assumptions if you only have 2 data points ! If you have many units of furniture, it's better and you could do a linear or nonlinear regression.
I guess I'd have to do a 3D graph, plotting completion time, NP & NF.
How do you do 3D regression? I guess I'd see a cluster but how do I quantify it? I only know the name Multivariate Analysis but evidently that's as far as it goes. . .
It's not 3D regression. It's multiiple regression. You have one response (time to assemble) and two predictors (number of panels, and number of fasteners). Is this an experiment you intend to do (if so, then you might want to rethink) ? If not, then the first thing to do is to produce some plots and see what the data looks like.
There may be a problem with multicolinearity.
It's so I can make more accurate bids for the times to assemble these furniture units, the idea being that I just look at the instruction sheet and can immediately estimate the expected & worst case time to some degree of confidence.
Seems like I have some multiple regression homework to do before I tackle this problem.
Last edited by Outlier; 03-18-2010 at 10:39 AM.
Posting Permissions
|
|