Noise reduction in prompt gamma spectra acquired in short times

Principal component analysis was applied for a noise reduction in prompt gamma spectra, especially acquired in short times, with the aim of improving signal-to-noise ratio. After the noise reduction, relevant information was preserved in the models while discarding much of the noise, and the signal-to-noise ratio was enhanced in the noisy prompt gamma spectra measured for 20 s. History data, which has multi elements and accurate peak positions of interest with a long enough measurement time, should be prepared for an effective noise reduction.