Figure (S1). Distribution Pattern of the Date Samples (Soft, Semi Dry ,Dry) in the 3-Dimensional

Figure (S1). Distribution pattern of the date samples (soft, semi dry ,dry) in the 3-dimensional PCA-based factor space of their transmittance FT-IR spectra.

(○,●) semi-dry, (□, ■) dry ( ,▼) soft:

(A)

(B)

Figure (S2).(A) Distribution pattern of the Kabkab dates samples in different geographical region in 2-dimension PCA plot.(○,●) Samal, (□, ■) Tangestan ( ,▼) Dashtestan: (B) Distribution pattern of the Zahedi dates samples in different geographical region in 3-dimension PCA-based factor space of their transmittance FT-IR spectra (○,●) Pesh Kooh, (□, ■) Posht Kooh.Open and filled markers denote prediction and calibratioon samples, respectiely.

Figure (S3).CV classification error for class (1,2 and 3) as a function of the latent variable PLS components used in the inner PLS on the transmittance FT-IR of datesamples.

Figure (S4).Distribution pattern of the date samples (soft-semi dry and dry ) in the PLS-DA factor spaces of their transmittance FT-IR spectra.

(○,●) semi-dry, (□, ■) dry ( ,▼) soft:

Figure (S5).Distribution pattern of the 12 date varietiesin the PLS-DA factor spaces of their transmittance FT-IR spectra.

Figure (S6).Number of misclassifications as a function of the number PLS components used in the inner PLS relation of ECVA on the transmittance FT-IR data.

Figure (S7). Distribution pattern of the date samples (12 varieties) in the 2-dimensional canonical variate space of their transmittance FT-IR spectra (whole region).

Figure (S8).Distribution pattern of the wavenumber values for the transmittance FT-IR data set obtained by 4-node self -organizing map.

Figure (S9). The extended canonical weights (for first three ones) obtained from CLoVA- ECVA method, using cluster (S4,4) from network size (n=4).

Figure (S10). Distribution pattern of the date samples (soft, semi –dry and dry) in the 2-dimensional canonical variate space of their transmittance FT-IR spectra(based on cluster (S4,4) of network size (n=4).)

Table (S1).The results of PCA application on the transmittance FT-IR data matrix of all date samples.

Table (S2).The percentage of misclassification obtained by different chemometrics methods for discrimination of date types.

Table (S3).The percentage of misclassification obtained by different chemometrics methods for discrimination of 12 date varieties