Prev: Saving images to create video
Next: fft of bearing stifness with respect to time
From: Greg Thom on 11 May 2010 06:54
Hai Matlab Users , specifically stats toolbox experts.

Here is my problem. A general statistics problem, which might be solved by matlab.

I have 5 measurements of the deviation ( from a baseline) of a temperature at 5 times of the day each day. I have made measurements for 50 days. So 5 * 50 measurements.

Now I need to compare these deviations for all the days and see if there is a "pattern" repeating accross number of days. I have been thinking of using slopes (calculated using five measurements per day) and then compare the slopes. What test can achieve this in matlab ?Is there a better way of comparing the means or slopes knowing that they have been computed from five sample points only ?

Thank you

Greg
From: Roger Stafford on 11 May 2010 17:20
"Greg Thom" <gregthom99(a)yahoo.com> wrote in message <hsbd0d$dnp$1(a)fred.mathworks.com>...
> Hai Matlab Users , specifically stats toolbox experts.
>
> Here is my problem. A general statistics problem, which might be solved by matlab.
>
> I have 5 measurements of the deviation ( from a baseline) of a temperature at 5 times of the day each day. I have made measurements for 50 days. So 5 * 50 measurements.
>
> Now I need to compare these deviations for all the days and see if there is a "pattern" repeating accross number of days. I have been thinking of using slopes (calculated using five measurements per day) and then compare the slopes. What test can achieve this in matlab ?Is there a better way of comparing the means or slopes knowing that they have been computed from five sample points only ?
>
> Thank you
>
> Greg

If you know the times of day the temperatures were taken at, this is just a problem in regression with time of day as the independent variable and temperature the dependent variable. Matlab is good at solving that kind of problem using the back slash operator ('\'). The "time" slope you obtain will be the best-fitting slope in a least squares sense.

Roger Stafford
 | 
Pages: 1
Prev: Saving images to create video
Next: fft of bearing stifness with respect to time