RESTORE:HBOOK:histograms
NOENT is the number of original entries in the histogram,
including underflows and overflows.
HLO is the sum of the weights for events below the histogram
lower limit. If you did not use weights when filling the
histogram, then HLO is the number of events below the histogram
lower limit.
HHI is the sum of the weights for events above the histogram
upper limit. If you did not use weights when filling the
histogram, then HHI is the number of events above the histogram
upper limit.
The number of equivalent events returned in HEQUIV is based on
the channel contents only. If w[i] represents the contents of
event i, then
number of equivalent events = (sum(w[i])^2/sum(w[i]^2)
This is what it is supposed to be, but... it seems to be just
the total number of entries in the histogram, which is the same
as NOENT. To calculate this value inside PHYSICA, you could use:
y = data[hstart[n]:hstart[n]+nbins[n]-1]
! hsum will be the sum of the weights inside the histogram limits
statistics\-message y hsum\sum
! hsum2 will be the sum of the squares of the weights
statistics y*y hsum2\sum
hequiv[n] = hsum*hsum/hsum2
The standard deviation returned in HSTD is based on the channel
contents only. If x[i] and w[i] represent the value and contents
of event i, and W = sum(w[i]), then
expectation value E(x) = sum(w[i]*x[i])/W
mean value = E(x)
central moment of order n, M(n) = E( (x - E(x))^n )
standard deviation = sqrt(M(2))
The mean value, or expectation value, returned in HMEAN is based
on the channel contents only. If x[i] and w[i] represent the value
and contents of event i, and W = sum(w[i]), then
mean value = sum(w[i]*x[i])/W