Electronic Theses and Dissertation

Penelitian ini bertujuan untuk mengetahui apakah metode Near Infrared (NIR) yang dikombinasikan dengan model klasifikasi kemometri Partial Least Square-Discriminant Analysis (PLS-DA) dan Support Vector Machine Classification (SVMC) mampu melihat perbedaan daun nilam varietas Lhokseumawe kontrol (PCLC), resisten (PCLR), dan terpapar jamur S. pogostemonis (PCLI) dan untuk melihat perbedaan secara morfologi daun nilam varietas Lhokseumawe kontrol, resisten, dan yang terpapar jamur S. pogostemonis. Klasifikasi spektrum NIR nilam varietas Lhokseumawe akibat terpapar jamur S. pogostemonis dianalisis menggunakan analisis multivariat kemometri dengan model klasifikasi Partial Least Square Discriminant Analysis (PLS-DA) dan Support Vector Machine Classification (SVMC). Klasifikasi secara kemometri menunjukkan bahwa tidak terdapat perbedaan antara nilam varietas Lhokseumawe kontrol, nilam resisten dan nilam terpapar jamur S. pogostemonis. Model klasifikasi PLS-DA memiliki score plot yang menunjukkan bahwa secara keseluruhan masing-masing kelompok sudah terklasifikasi cukup bagus ditunjukkan dengan nilai kalibrasi data NIR dengan nilai R2 dan Root Mean Standard Error of Calibration (RMSEC) masing-masing sebesar 0,996 dan 0,04. Selanjutnya untuk nilai R2 validasi dan RMSEC sebesar 0,994 dan 0,04. Model klasifikasi SVMC menghasilkan training accuracy sebesar 72,25% dengan cross-validation accuracy-nya sebesar 65,80%. Hasil penelitian ini menunjukkan bahwa data spektrum NIR yang dikombinasikan dengan kemometri dapat digunakan untuk mengklasifikasikan sampel daun nilam kontrol, resisten dan yang terpapar S. pogostemonis. Selanjutnya, hasil analisis SEM pada sampel daun juga menunjukkan adanya perbedaan secara morfologi pada nilam yang terpapar.

Kata kunci: Pogostemon cablin, Synchytrium pogostemonis, NIR, PCA, PLS-DA, SVMC, SEM

This study aims to determine whether the NIR method combined with Partial Least Square-Discriminant Analysis (PLS-DA) and Support Vector Machine Classification (SVMC) chemometrics classification model is able to see differences in patchouli leaves of Lhokseumawe varieties of control (PCLC), resistant (PCLR), and infected by S. pogostemonis fungus (PCLI) and to see morphological differences in patchouli leaves of Lhokseumawe varieties of control, resistant, and infected by S. pogostemonis fungus. Classification of the NIR spectrum of Lhokseumawe patchouli due to infected by the S. pogostemonis fungus was analyzed using chemometric multivariate analysis with Partial Least Square Discriminant Analysis (PLS-DA) and Support Vector Machine Classification (SVMC) classification models. Chemometric classification showed that there was no difference between the control, resistant and infected by S. pogostemonis patchouli varieties of Lhokseumawe. The PLS-DA classification model has a score plot that shows that overall each group has been classified quite well as indicated by the calibration value of NIR data with R2 and Root Mean Standard Error of Calibration (RMSEC) values obtained are 0.996 and 0.04, respectively. Furthermore, the validation R2 and RMSEC values are 0.994 and 0.04. The SVMC classification model showed that training accuracy value obtained is 72.25% with the cross-validation accuracy value is 65.80%. The results of this study indicate that NIR spectrum data combined with chemometrics can be used to classify control, resistant and infected patchouli leaf samples of S. pogostemonis. Furthermore, the results of SEM analysis on leaf samples also showed morphological differences on infected patchouli. Keyword: Pogostemon cablin, Synchytrium pogostemonis, NIR, PCA, PLS-DA, SVMC, SEM