Page 1 of 1

HVSR statistical comparison

Posted: Thu Mar 22, 2012 6:34 pm
by riccardo
Hi mark, excuse me but I haveanother question.
This question is about the tool Statistical comparison for HVSR

for example:
Frequency peak statistics:
Reference: 0.16 0.18 0.19 Hz Test: 0.14 0.17 0.19 Hz
F-Test (0.95): 0.526946 in [ 0.901066 , 0.947927 ] => NOT similar variances
Student Test (0.95, equal variances): |delta(f0)|=0.0136126 > 0.00971345 => NOT similar peaks
Student Test (0.95, unequal variances): |delta(f0)|=0.0136126 > 0.00771445 => NOT similar peaks

Reference are the picks with St. dev. of H/V ratio and this it's Ok

F-test (0,95) is the confidence that I can chose for my data;
what is: 0.526946 in [ 0.901066, 0.947927]

I know what is F-test but I can't understand what are that numbers

can you help me?

Re: HVSR statistical comparison

Posted: Tue Apr 10, 2012 12:24 pm
by admin
The "Test" values must fall within a range to satisfy a 95% confidence test. In your case, the variance of 0.52 for the test is out of the expected range computed from the reference peak.
The comments are related only to the statistical comparison of the peak statistics (not the complete H/V curve).