Putting the Spectrum in Bins (Data Reduction)
For multivariate statistical analysis, it is often useful to replace the spectrum with a smaller set of representative values. Binning reduces the number of variables in a spectrum, making it easier to process with statistical software. What these programs typically require is a set of integrals calculated (over the whole spectrum or the central part of it) at fixed, regular, narrow intervals. The resulting binned spectrum is displayed as a standard spectrum and resembles the original one. Both spectra can be exported in tabular form (as ASCII files) with the command File > Export.
How to Bin a Spectrum
- Process the spectrum as usual and reference it against TMS or another signal.
- Choose Process > Make Bins. For each frequency axis (if the spectrum has more than one dimension) specify:
- The lowest frequency, in ppm, of the region of interest.
- The highest frequency of the region of interest.
- The size of each bin, in ppm.
- Normally you set the last parameter (edge corr.) to zero, which means it is not used. This parameter performs an additional baseline correction. In most cases, this extra correction is unnecessary because the baseline should already have been corrected before binning. In practice, if you specify a value n for edge corr., iNMR calculates the average value of the first n and the last n bins and subtracts it from all the bins.
After binning, some commands are disabled to keep the integrity of the data. For example, it is not possible to change the scale reference.