Malaysian Journal of Analytical Sciences Vol 26 No 2 (2022): 295 - 302

 

 

 

 

KAFFIR LIME OIL QUALITY GRADING USING NON-LINEAR SUPPORT VECTOR MACHINE WITH DIFFERENT KERNELS

 

(Penentuan Kualiti Minyak Limau Purut Dengan Menggunakan Mesin Vektor Sokongan Bukan Linear Dengan Kernel Berbeza)

 

Nor Syahira Jak Jailani¹*, Zuraida Muhammad², Nor Salwa Damanhuri², Muhd Hezri Fazalul Rahiman¹, Mohd Nasir Taib¹

 

¹College of Engineering,

Universiti Teknologi MARA, 40450 Shah Alam, Selangor, Malaysia

²College of Engineering,

Universiti Teknologi MARA,

13500 Permatang Pauh, Pulau Pinang, Malaysia

 

*Corresponding author:  norsyahira_jakjailani@yahoo.com

 

 

Received: 23 September 2021; Accepted:  5 February 2022; Published:  28 April 2022

 

 

Abstract

 

Nowadays, Kaffir lime oil is one of the highly demand in the industries and sold with various market prices. But sometime the most expensive price of kaffir lime oil does not guarantee the best quality of this oil. Currently, kaffir lime oil quality grading is based only on human sensory evaluation such as smell and visual which accompanying with confusion and inconsistent result. This is due to human sensory evaluations have limitations such as easily fatigue and unable to manage mass products of the oil in time. This paper presents the classification of significant chemical compound of kaffir lime oil in the oil quality grading by using Non-Linear Support Vector Machine (NSVM). The objective of this study is to classify the quality of kaffir lime oil whether it is high or low in quality only by tuning the two different kernels into NSVM. This project has used up to 90 samples of kaffir lime oil data from various high to low quality to prepare NSVM with two different kernels. The implementation of this project was performed by using MATLAB version R2020A. The result appeared NSVM model with RBF kernel is better than NSVM model with Polynomial kernel. It was discovered that RBF kernel was able to generate 100% accuracy, specificity, precision, and sensitivity compared to Polynomial kernel. Since the finding and outcome were effective and significant, thus this study contributes a lot of benefits for future study especially in kaffir lime field.

 

Keywords:   non-linear support vector machine, radial basis function, polynomial, kaffir lime oil, high/low quality

 

Abstrak

Saat ini, minyak limau purut mendapat permintaan tinggi di industri, dan dijual dengan harga yang berbeda di pasar. Tetapi harga minyak limau purut yang mahal tidak menjamin kualiti tinggi minyak ini. Pada masa ini, penilaian kualiti minyak limau purut berdasarkan penilaian deria seperti hidung dan mata menimbulkan kekeliruan dan hasil yang tidak konsisten. Ini kerana penilaian deria mempunyai batasan dan mudah letih dan tidak dapat menangani banyak sampel sekaligus. Oleh itu, kajian inimembentangkan klasifikasi sebatian kimia yang penting dalam minyak limau purut untuk gred kualiti minyak dengan menggunakan Mesin Vektor Sokongan Bukan Linear (NSVM). Objektif kajian ini adalah untuk mengklasifikasikan kualiti minyak limau purut sama ada berkualiti tinggi atau rendah dengan menyesuaikan dua kernel yang berbeza ke dalam NSVM. Projek ini menggunakan 90 sampel data minyak limau purut dari berkualiti tinggi hingga rendah untuk melatih NSVM dengan kernel yang berbeza. Pelaksanaan projek ini dilakukan dengan menggunakan MATLAB versi R2020A. Hasilnya menunjukkan model NSVM dengan kernel RBF lebih baik daripada model NSVM dengan kernel Polinomial. Didapati kernel RBF mampu menghasilkan ketepatan, kekhususan, ketepatan, dan kepekaan 100% berbanding dengan kernel Polinomial. Penemuan dan hasil ini berkesan dan signifikan sekali gus akan mendorong dan memberi banyak faedah untuk kajian masa depan terutama dalam bidag limau purut

 

Kata kunci:  mesin vektor sokongan bukan linear, fungsi asas radial, polinomial, minyak limau purut, kualiti tinggi/rendah

 

 

 

Graphical Abstract

 

 

 

References

1.      Abas, M. A., Zubir, N. S. A., Ismail, N., Yassin, I. M., Ali, N. A. M., Rahiman, M. H. F. and Taib, M. N. (2017). Agarwood oil quality classifier using machine learning. Journal of Fundamental and Applied Sciences, 9(4): 62-76.

2.      Haron, M. H., Taib, M. N., Ismail, N., Mohd Ali, N. A. and Tajuddin, S. N. (2019). Statistical analysis of agarwood oil compounds based on GC-MS data. 9^th IEEE Control System Grad. Research Colloquium: pp. 27-30.

3.      Azah, M. N., Husni, S. S., Mailina, J., Sahrim, L., Majid, J. A. and Faridz, Z. M. (2013). Classification of agarwood (gaharu) by resin content. Journal of Tropical Forest Science, 2013: 213-219.

4.      Ismail, N., Azah, M. A. N., Jamil, M., Rahiman, M. H. F., Tajuddin, S. N. and Taib, M. N. (2013). Analysis of high quality agarwood oil chemical compounds by means of SPME/GC-MS and Z-score technique. Malaysian Journal of Analytical Sciences, 17(3): 403-413.

5.      Jantan, H., Yassin, I. M., Zabidi, A., Ismail, N. and Ali, M. M. (2017). Differentiation of agarwood oil quality using support vector machine (SVM). Journal Engineering Applied Sciences, 12(15): 3810-3812.

6.      Mohd Sauid, S. and Aswandi, F. A. (2018). Extraction methods of essential oil from kaffir lime (Citrus hystrix): A review. Malaysian Journal of Chemical Engineering and Technology, 1: 56-64.

7.      Kasuan, N., Muhammad, Z., Yusoff, Z., Fazalul Rahiman, M. H., Taib, M. N. and Haiyee, Z. A. (2013). Extraction of Citrus hystrix DC (kaffir lime) essential oil using automated steam distillation process: analysis of volatile compounds. Malaysian Journal of Analytical Sciences, 17: 359-369.

8.      Haiyee Z. A. and Winitkitcharoen, C. (2012). Extraction of volatile oil from kaffir lime leaves (Citrus hystrix) using pressurize liquid extraction. International Journal Food, Nutrition Public Health, 5: 201-210.

9.      Muhammad, Z., Yusoff, Z. M. and Nordin, M. N. N. (2013). Steam Distillation with induction heating system: analysis of kaffir lime oil compound and production yield at various temperatures. Malaysian Journal of Analytical Sciences, 17(2): 340-347.

10.   Ngan, T. T. K., Hien, T. T., Tien, L. X., Anh, T. T., Quan, P. M., Cang, M. H. and Toan, T. Q. (2019). Physico-chemical profile of essential oil of kaffir lime (Citrus hystrix DC) grown in An Giang Province, Vietnam. Asian Journal Chemistry, 31(12): 2855-2858.

11.   Hongatanaworkit, T. and Buchbauer, G. (2007). Chemical composition and stimuating effect of Citrus hystrix oil on human. Flavour Fragrance Journal, 22(4): 443-449.

12.   Dewi, C. and Basuki, A. (2018). Identifying citronella plants from UAV imagery using support vector machine. Telkomnika, 16(4): 1877?1885.

13.   Windha Mega, P. D. and Haryoko (2019). Optimization of parameter support vector machine (SVM) using genetic algorithm to review go-jek?s services. 4^th International Conference Information Technology, Information System, and Electrical Engineering, 6: 301-304.

14.   Zakiah, M. Y., Zuraida, M., Nurhani, K., Mohd, H. F. R. and Mohd, N. T. (2013). Effect of temperature on kaffir lime oil by using hydro-diffusion steam distillation system. Malaysian Journal of Analytical Sciences, 17(2), 326-339.

15.   Lubinska-Szczygieł, M., R?żańska, A., Namieśnik, J., Dymerski, T., Shafreen, R. B., Weisz, M., ... & Gorinstein, S. (2018). Quality of limes juices based on the aroma and antioxidant properties. Food Control, 89: 270-279.

16.   An, T. N. T., Ngan, T. T. K., Van, C. K., Anh, H. L. T., Minh, L. V. and Ay, N. V. (2021, March). The major and minor components of Kaffir Lime (Citrus hystrix DC) essential oil in the steam distillation process. In IOP Conference Series: Materials Science and Engineering, 1092(1): 012082.

17.   Jailani, N. S. J., Muhammad, Z., Rahiman, M. H. F. and Taib, M. N. (2021). Analysis of kaffir lime oil chemical compounds by gas chromatography-mass spectrometry (GC-MS) and Z-Score technique. In 2021 IEEE International Conference on Automatic Control & Intelligent Systems: pp. 110-113.

18.   Ismail, N., Rahiman, M. H. F., Taib, M. N., Ali, N. A. M., Jamil, M. and Tajuddin, S. N. (2014). Application of ANN in agarwood oil grade classification. In 2014 IEEE 10th International Colloquium on Signal Processing and its Applications: pp. 216-220.

19.   Sevgi, L. (2007). Synthetic radar-Signal environment: Computer generation of signal, noise, and clutter. IEEE Antennas and Propagation Magazine, 49(5): 192-198.

20.   Ismail, N. S., Ismail, N., Rahiman, M. H. F., Taib, M. N., Ali, N. A. M. and Tajuddin, S. N. (2018, October). Polynomial tuned kernel parameter in SVM of agarwood oil for quality classification. In 2018 IEEE International Conference on Automatic Control and Intelligent Systems: pp. 77-82.

21.   Ariffin, B. M., Ismail, N., Ali, N. A. M., Rahiman, M. H. F., Tajuddin, S. N. and Taib, M. N. (2018). Multilayer perceptron (MLP) tuned kernel parameter in support vector machine (SVM) for agarwood oil compounds quality classification. J. Tomography System and Sensors Application, 1(1): 1-15.

22.   Raif, M. A. A., Ismail, N., Ali, N. A. M., Rahiman, M. H. F., Tajuddin, S. N. and Taib, M. N. (2019). Quadratic tuned kernel parameter in Non-linear support vector machine (SVM) for agarwood oil compounds quality classification. Indonesia Journal Electronic Engineering Computer Science, 17(3): 1371-1376.

23.   Patle, A. and Chouhan, D. S. (2013). SVM kernel functions for classification. In 2013 International Conference on Advances in Technology and Engineering, pp. 1-9.

24.   Apostolidis-Afentoulis, V. and Lioufi, K. I. (2015). SVM classification with linear and RBF kernels: pp. 1-7.

25.   Kamarulzaini, K. A. A., Ismail, N., Rahiman, M. H. F., Taib, M. N., Ali, N. A. M. and Tajuddin, S. N. (2018). Evaluation of RBF and MLP in SVM kernel tuned parameters for agarwood oil quality classification. In 2018 IEEE 14th International Colloquium on Signal Processing & Its Applications, pp. 250-254.