Abstract: A Fourier transform (FT) was used as a tool to reduce the number of variables in pattern recognition. It is carried out in the frequency domain starting from the first 15 Fourier coefficients (FFT coefficients) as the input to the PLS(partical least square) method and the PCR(principal-component regression) method. Comparison between the PLS FFT method and other calibration methods applied in the original domain (wavelength domain WD) and in the FFT domain. The results demonstrate that FT is a fast and powerful reduction method. This also shows that the method can be used to investigate the structure of the data.

Key words: Fourier coefficients, feature selection, apple sugar content, NIR