Malaysian Journal of Analytical Sciences, Vol 26 No 6 (2022): 1240 – 1248

 

THE APPLICATION OF ELECTRONIC NOSE COUPLED WITH 80:20 K-NEAREST NEIGHBORS CLASSIFICATION TECHNIQUE FOR AGARWOOD OIL QUALITY INDEX ESTABLISHMENT

 

(Kombinasi Aplikasi Hidung Elektronik Bersama Teknik Klasifikasi 80:20 K-Nearest Neighbors Untuk Pembangunan Indek Kualiti Minyak Gaharu)

 

Sahrim Lias1, 2, Abd Majid Jalil2, Mailina Jamil2, Mohd Shafik Yuzman Tolmanan2, Nurlaila Ismail1,

Mohd Nasir Taib1

 

1 Faculty of Electrical Engineering,

University Teknologi Mara (UiTM), Shah Alam 40000, Selangor, Malaysia

2 Natural Products Division,

Forest Research Institute Malaysia (FRIM), Kepong 52109, Selangor, Malaysia 

 

 *Corresponding author: sahrim@frim.gov.my

 

 

Received: 3 February 2022; Accepted: 19 May 2022; Published:  27 December 2022

 

 

Abstract

Agarwood oil is highly valued with numerous uses and benefits to the user. However, traditional agarwood grading techniques are subjective, time consuming and labor intensive with poorly reproducible results, in our study, a commercial Fox4000 electronic nose (EN) with AlphaSOFT software was evaluated as a substitute for human nose sniff assessment. Eight pure agarwood oil samples were used: the JBD sample that was selected as the initial high-grade reference sample based on previous research, and seven samples new to the FRIM collection. All samples were subjected to the EN tests and the seven new samples compared with the JBD sample using the non-parametric Wilcoxon’s signed-rank test. The REG sample that was closest in aroma profile to JBD was then chosen as the high-grade reference sample for subsequent tests. Varying volumes of REG were blended with gurjun balsam and sandalwood pure oils to produce nineteen blended samples containing increasing percentages of REG. The 19 samples were then used in the EN tests for database establishment. All 19 blended samples were subjected to the EN tests and resulting data transformed using a z-score scale to generate an Agarwood Oil Quality Index (AOQI) that was classified into three levels based on percentage volume of agarwood oil in the oil blend. The EN significant sensor selection process was conducted using Spearman’s rho correlation (SRC) and stepwise multi linear regression (SMLR). Finally, the 80:20 k-Nearest Neighbors (kNN) classifier was used to evaluate the AOQI model with the confusion-matrix based performance measure (CMBPM) as the classifier performance evaluator. The results showed SRC with 11 selected sensors outperformed SMLR with 80:20 k-NN test accuracy equal to 89.5%.

 

Keywords: agarwood oil grading, electronic nose, 80:20 k-NN classifier

 

 

Abstrak

Minyak gaharu dikenali mempunyai harga yang tinggi dengan pelbagai kegunaan dan manfaat kepada pengguna. Walau bagaimanapun, teknik penggredan gaharu secara tradisional adalah sangat terhad dari segi subjektiviti, kebolehulangan yang rendah, penggunaan masa, dankos buruh yang mahal. Dalam kajian ini,  Hidung Elektronik (EN) Fox4000 komersial dengan perisian AlphaSOFT telah digunakan bagi menggantikan penilaian menghidu menggunakan hidung manusia. Lapan sampel minyak gaharu tulen telah digunakan. Sampel JBD telah dipilih sebagai sampel rujukan bergred tinggi berdasarkan kajian terdahulu . Semua sampel tertakluk kepada eksperimen EN dan dibandingkan dengan sampel JBD menggunakan ujian peringkat-bertanda Wilcoxon bukan parametrik  untuk pemilihan sampel rujukan gred tinggi yang baharu. Sampel REG telah dipilih dan dicampur dengan sampel minyak keruing dan cendana tulen . Kesemua 19 sampel campuran telah melalui ujian EN untuk penubuhan tiga tahap utama  Indeks Kualiti Minyak Gaharu (AOQI) menggunakan skala skor-z. Proses pemilihan sensor EN yang signifikan telah dijalankan menggunakan teknik Spearman Rho Correlation (SRC) dan Stepwise Multi Linear Regression (SMLR). Akhirnya, teknik kelasifikasi 80:20 k-Nearest Neighbors (k-NN) telah digunakan untuk menilai model AOQI dengan teknik Confusion-Matrix Based Performance Measure (CMBPM) sebagai penilai prestasi pengelasan. Keputusan menunjukkan SRC dengan 11 sensor terpilih mengatasi prestasi SMLR dengan ketepatan ujian 80:20 k-NN bersamaan dengan 89.50%.

 

Kata kunci: penggredan minyak gaharu, hidung elektronik, pengelasan 80:20 k-NN

 

 

Graphical Abstract

 

References

1.         Yahya, A. Z., Wai Mun, H. and Hassan, N. H. (2018). KARAS (Aquilaria malaccensis) A plantation species in Peninsular Malaysia. Forest Research Institute Malaysia (FRIM). Kuala Lumpur, Malaysia.

2.         Muhammad, I., Jusoh, M. Z. and Sajap, A. S. (2018). Gaharu Komoditi Ketiga Negara. University Putra Malaysia (UPM). Selangor, Malaysia.

3.         Forestry Department Peninsular Malaysia (2015). Manual in grading of agarwood. Forestry Department Peninsular Malaysia (FDPM). Selangor, Malaysia.

4.         Malaysian Timber industry Board (2014). Panduan pengkelasan produk gaharu Malaysia. Malaysian Timber industry Board (MTIB). Kuala Lumpur, Malaysia.

5.         Lias, Sahrim, Ali, N. A. M., Jamil, M., Jalil, A. M. and Othman, M. F. (2016). Discrimination of pure and mixture agarwood oils via electronic nose coupled with k-NN kfold classifier. Procedia Chemistry, 20: 63-68.

6.         Ismail, N., Rahiman, M. H. F., Taib, M. N., Ali, N. A. M., Jamil, M. and Tajuddin, S. N. (2013). Classification of the quality of agarwood oils from Malaysia using Z-score technique. Proceedings - 2013 IEEE 3rd International Conference on System Engineering and Technology, 2013: 78-82.

7.         Haslina, E., Latib, A., Najib, M. S., Mohd, C., Che, A. and Tajuddin, S. N. (2018). Analysis of different quality agarwood oil (Aquilaria malaccensis) and sensory study. Journal of Telecommunication, Electronic and Computer Engineering, 10(1): 57-61.

8.         Zubir, N. S. A., Abas, M. A., Ismail, N., Ali, N. A. M., Rahiman, M. H. F., Mun, N. K., Tajuddin, S. H. and Taib, M. N. (2017). Analysis of algorithms variation in multilayer perceptron neural network for agarwood oil qualities classification. 8th Control and System Graduate Research Colloquium, ICSGRC 2017 - Proceedings: 122-126.

9.         Ismail, N., Rahiman, M. H. F., Taib, M. N., Ali, N. A. M., Jamil, M. and Tajuddin, S. N. (2013). Analysis of chemical compounds of agarwood oil based on headspace-solid phase microextraction combined with gas chromatography mass-spectrometry. 9th International Colloquium on Signal Processing and Its Applications: 215-218.

10.      Ismail, S. N., Maulidiani, M., Akhtar, M. T., Abas, F., Ismail, I. S., Khatib, A., Mohamad Ali, N. A. and   Shaari, K. (2017). Discriminative analysis of different grades of gaharu (Aquilaria malaccensis Lamk.) via1H-NMR-based metabolomics using PLS-DA and random forests classification models. Molecules, 22(10): 1612.

11.      Lias, S., Mohamad Ali, N. A., Jamil, M., Tolmanan, M. S. Y. and Misman, M. A. (2018). A study on the application of electronic nose coupled with dfa and statistical analysis for evaluating the relationship between sample volumes versus sensor intensity of agarwood essential oils blending ratio. MATEC Web of Conferences, 2018: 201-02008.

12.      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. 10th International Colloquium on Signal Processing and Its Applications, CSPA 2014: 216-220.


13.      Ismail, N., Rahiman, M. H. F., Taib, M. N., Ali, N. A. M., Jamil, M. and Tajuddin, S. N. (2014). A review study of agarwood oil and its quality analysis, Jurnal Teknologi, 1: 37-42.

14.      Karimifard, S., Ahmadian, A., Khoshnevisan, M. and Nambakhsh, M. S. (2006). Morphological heart arrhythmia detection using hermitian basis functions and kNN classifier. Annual International Conference of the IEEE Engineering in Medicine and Biology - Proceedings, (4): 1367-1370.

15.      Liang, K. and Wang, D. (2019). KNN _ SVM based information transmission method for renewable energy power generation. IEEE Sustainable Power and Energy Conference, 2019: 1671–1676.

16.      Wasule, V. and Sonar, P. (2017). Classification of brain MRI using SVM and KNN classifier. 3rd IEEE International Conference on Sensing, Signal Processing and Security, 2017: 218-223.

17.      Bharate, A. A. and Shirdhonkar, M. S. (2020). Classification of grape leaves using KNN and SVM classifiers. 4th International Conference on Computing Methodologies and Communication, 2020: 745-749.

18.      Xu, M., Wang, J. and Gu, S. (2019). Rapid identification of tea quality by E-nose and computer vision combining with a synergetic data fusion strategy. Journal of Food Engineering, 241: 10-17.