Page 86 - BOOK OF PROGRAM ICAGRI 2

#3419

Comparison of principle component and partial least
square regression method in NIRS data analysis for cocoa

bean quality assessment

1
1
M Kamal , M I Sulaiman , A A Munawar 2
1 Department of Agriculture Product Technology, Universitas Syiah Kuala,
Banda Aceh, Indonesia
2 Department of Agricultural Engineering, Universitas Syiah Kuala, Banda
Aceh, Indonesia

E-mail : ikhsan.sulaiman@unsyiah.ac.id

Abstract. In general, determining the quality attributes of fermented cocoa beans must go
through two stages: (1) split test of cocoa beans to visually see the degree of dryness and aroma
resulting from the fermentation process, (2) to obtain data on the quality attributes of cocoa
beans such as moisture content and fat content, it is usually carried out through a chemical
laboratory test stage and takes time. Therefore, the development of NIRS instrument, it is hoped
to predict quality attributes of fermented cocoa beans without going through the cleavage stage
and can save time to obtain data on the quality of cocoa beans. This study aims to predict and
compare the results of cocoa beans quality tests by using destructive methods (chemical
laboratories) and with using NIRS instrument to determine including tests of moisture content
and fat content which are attributes of the quality of cocoa beans. The results showed absorption
peaks in the moisture content spectrum of cocoa beans were obtained at a wavelength of 1497-
1536 nm, the peak fat content was obtained at a wavelength of 2120-2292 nm. By using two
regression methods (PCR and PLSR), the R-square value obtained by the PCR regression
method was 0.83 (83%) higher than PLSR= 0.76 (76%), and the RMSE indeks obtained in the
PCR regression method = 0,51 is lower than that of PLSR = 0.61. In the analysis of fat content,
the coefficient of determination (R2) PLSR was 0.77 greater than that of PCR. It can be
interpreted that the PLSR regression method is able to predict fat content better with only 23%
deviation.

© 2020 ICAGRI - The 2nd International Conference on Agriculture and Bioindustry (ICAGRI), 83
Agriculture Faculty, Universitas Syiah Kuala. All Rights Reserved