感測器漂移是一個損害電子鼻系統可靠性的重要問題，感測器在實際應用中長時間的運行而導致的老化問題以及環境變化等因素都會使感測器的反應發生變化，嚴重降低感測器的使用壽命。而漂移問題也意味著數據分佈發生了變化，使原本的數據資料或分類模型不再適用，進而導致降低分類模型對氣體的辨識率。
本文提出了一種可遷移特徵的分類器結合特徵加權來解決傳感器漂移的方法。根據特徵的性質：分離性和遷移性，計算出每種特徵的特徵分數，並根據特徵分數選出最佳的特徵子集。此外，遷移性的分數將用作分類器特徵加權的權重，以考慮每種特徵的遷移性對分類器的影響。
最後，在分類器方面，為了減少源域以及目標域之間的分佈差異，提出遷移極限學習機(TELM)，以實現跨域知識轉移(Cross-Domain Knowledge Transferring)，達到知識轉移的功能進而增加分類器在不同域之間的轉移能力。
本研究利用紀錄感測器漂移問題的數據來驗證可遷移特徵的分類器之性能，實驗結果表明，採用此方法可以提高漂移數據集的平均分類準確度。

Sensor drift is a significant problem that damages the reliability of the electronic nose system(E-nose). Factors such as aging problems and environmental changes caused by the long-term operation of the sensor in practical applications will change the sensor's response and seriously reduce the sensor's service life. The drift problem also means that the data distribution has changed so that the original data or classification model is no longer applicable, reducing the classification model's recognition rate for the gas.
This paper proposes a transferable feature-based classifier combined with feature weighting to address sensor drift. The feature score of each feature is calculated according to the separability and transferability of feature, and the best feature subset is selected according to the feature score. Additionally, the transferability score will be used as a weight for feature weighting to consider the effect of each feature's transferability on the classifier.
Finally, in terms of classifiers, to reduce the distribution difference between domains (source domain and target domain), Transfer Extreme Learning Machine (TELM) is proposed to realize cross-domain knowledge transfer to achieve knowledge transfer, increasing the transferability of the classifier between different domains.
This study uses data recorded on sensor drift problems to verify the classifier's performance with transferable features. The experimental results show that this method can improve the average classification accuracy of drift datasets.

摘要 i
ABSTRACT ii
目錄 iii
圖目錄 v
表目錄 vii
第 1 章 緒論 1
1.1. 嗅覺機制 1
1.2. 電子鼻系統 (Electronic Nose System) 1
1.3. 章節介紹 4
第 2 章 文獻回顧 5
2.1. 電子鼻系統的特徵擷取 (Feature Extraction) 5
2.2. 電子鼻系統的感測器漂移問題 6
2.2.1. 特徵層面 (Feature-Level) 7
2.2.2. 決策層面 (Decision-Level) 12
2.3. 研究動機 17
第 3 章 可遷移特徵的分類器 (Transferable Feature-Based Classifier) 18
3.1. SM-TM特徵選擇 (SM-TM Feature Selection) 19
3.1.1. 類別分離性 (Separability Metrics, SM) 19
3.1.2. 可遷移性 (Transferability Metrics, TM) 21
3.1.3. 特徵分數 (Feature Score) 23
3.2. 遷移極限學習機 (Transfer Extreme Learning Machine, TELM) 24
3.2.1. 極限學習機 (Extreme Learning Machine, ELM) 24
3.2.2. 遷移極限學習機器分類器 (Transfer Extreme Learning Machine Classifier, TELM) 26
3.3. 極限學習機分類器的加權特徵 (Weighted Feature for Extreme Learning Machine Classifier) 27
第 4 章 實驗結果與討論 29
4.1. 實驗數據集介紹 29
4.2. 實驗結果與討論 31
4.2.1. 遷移樣本(Transfer Samples)的選擇 31
4.2.2. 結果比較 31
4.2.3. 分類器參數探討 41
第 5 章 結論與未來發展 43
參考文獻 44

[1] Gardner, J. W. and Bartlett, P. N. A brief history of electronic noses. Sensors and
Actuators B: Chemical, vol. 18, no. 1-3, pp. 210-211, 1994.
[2] Axel, R. and Buck, L. B. Odorant receptors and the organization of the olfactory system. Nobel Prize in Physiology or Medicine, 2004.
[3] Karakaya, D., Ulucan, O., Turkan, M. Electronic Nose and Its Applications: A Survey. Int. J. Autom. Comput. 17, pp. 179-209, 2020.
[4] Loutfi, A., Coradeschi, S., Mani, G. K., Shankar, P., Rayappan, J. B. B. Electronic noses for food quality: A review. Journal of Food Engineering, vol. 144, pp. 103-111, 2015.
[5] Compagnone, D., Faieta, M., Pizzoni, D., Natale, C. D., Paolesse, R., Van Caelenberg, T., Beheydt, B., Pittia, P. Quartz crystal microbalance gas sensor arrays for the quality control of chocolate. Sensors and Actuators B: Chemical, vol. 207, pp. 1114-1120, 2015.
[6] Dutta, R., Hines, E. L., Gardner, J. W., Kashwan, K. R., Bhuyan, M. Tea quality prediction using a tin oxide-based electronic nose: An artificial intelligence approach. Sensors and Actuators B: Chemical, vol. 94, no. 2, pp. 228-237, 2003.
[7] Wilson, A. D. Electronic-nose devices-potential for noninvasive early disease-detection applications. Annals of Clinical Case Reports, vol.2, Article number 1401, 2017.
[8] Wong, D. M., Fang, C. Y., Chen, L. Y., Chiu, C. I., Chou, T. I., Wu, C. C., Chiu, S. W., Tang, K. T. Development of a breath detection method based E-nose system for lung cancer identification. In Proceedings of IEEE International Conference on Applied System Invention, IEEE, Chiba, Japan, pp. 1119-1120, 2018.
[9] Mumyakmaz, B. and Karabacak, K. An E-nose-based indoor air quality monitoring system: Prediction of combustible and toxic gas concentrations. Turkish Journal of Electrical Engineering & Computer Sciences, vol. 23, no. 3, pp. 729-740, 2015.
[10] Wang, J. and Yinon, J. Electrochemical sensing of explosives. In Counterterrorist Detection Techniques of Explosives. Amsterdam, Netherlands: Elsevier, pp. 91-107, 2007.
[11] Estakhroueiyeh, H. R. and Rashedi, E. Detecting moldy Bread using an E-nose and the KNN classifier. 2015 5th International Conference on Computer and Knowledge Engineering (ICCKE), pp. 251-255, 2015.
[12] Gaudioso, M., Khalaf, W., Pace, C. On the Use of the SVM Approach in Analyzing an Electronic Nose. 7th International Conference on Hybrid Intelligent Systems (HIS 2007), pp. 42-46, 2007.
[13] Matteo, P. and Giorgio, S. Classification of electronic nose data with support vector machines. Sensors and Actuators B: Chemical, vol. 107, Issue 2, pp. 730-737, 2005.
[14] Peng, P., Zhao, X., Pan, X., Ye, W. Gas classification using deep convolutional neural networks. Sensors 18, no. 157, 2018.
[15] Wei, G., Li, G., Zhao, J., He, A. Development of a LeNet-5 gas identification CNN structure for electronic noses. Sensors 19, no. 1:217, 2019.
[16] Zhang, L. and Zhang, D. Efficient solutions for discreteness, drift, and disturbance (3d) in electronic olfaction. IEEE Transactions on Systems, Man, and Cybernetics: Systems, vol. 48, no. 2, pp. 242-254, 2016.
[17] Yan, J., Guo, X., Duan, S., Jia, P., Wang, L., Peng, C., Zhang, S. Electronic Nose Feature Extraction Methods: A Review. Sensors 15, no. 11, pp. 27804-27831, 2 Nov. 2015.
[18] Artursson, T., Eklöv, T., Lundström, I., Mårtensson, P., Sjöström, M., Holmberg, M. Drift correction for gas sensors using multivariate methods. J. Chemometrics, vol. 14, no. 5/6, pp. 711-723, 2000.
[19] Padilla, M., Perera, A., Montoliu, I., Chaudry, A., Persaud, K., Marco, S. Drift compensation of gas sensor array data by orthogonal signal correction. Chemometrics Intell. Laboratory Syst., vol. 100, no. 1, pp. 28-35, 2010.
[20] Zhang, L., Liu, Y., He, Z., Liu, J., Deng, P., Zhou, X. Anti-drift in e-nose: A subspace projection approach with drift reduction. Sensors Actuators B: Chem., vol. 253, pp. 407-417, Jun. 2017.
[21] Vergara, A., Vembu, S., Ayhan, T., Ryan, M. A., Homer, M. L., Huerta, R. Chemical gas sensor drift compensation using classifier ensembles. Sensors Actuators B: Chem., vol. 166, pp. 320-329, 2012.
[22] Zhang, L. and Zhang, D. Domain adaptation extreme learning machines for drift compensation in e-nose systems. IEEE Trans. Instrume. Meas., vol. 64, no. 7, pp. 1790-1801, Jul. 2015.
[23] Yi, R., Yan, J., Shi, D., Tian, Y., Chen, F., Wang, Z., Duan, S. Improving the performance of drifted/shifted electronic nose systems by cross-domain transfer using common transfer samples. Sensors and Actuators B: Chemical, vol. 329, 2021.
[24] Huang, G., Song, S., Gupta, J. N. D., Wu, C. Semi-supervised and unsupervised extreme learning machines. IEEE Transactions on Cybernetics, vol. 44, no. 12, pp. 2405-2417, Dec. 2014.
[25] Sharifmoghadam, M. and Jazayeriy, H. Breast cancer classification using AdaBoost-extreme learning machine. in Proceeding of Iranian Conference on Signal Processing and Intelligent Systems (ICSPIS 2019), pp. 1-8, IEEE, Shahrood, Iran, Dec. 2019.
[26] Borgwardt, K. M., Gretton, A., Rasch, M. J., Kriegel, H. P., Schölkopf, B., Smola, J. Integrating structured biological data by Kernel Maximum Mean Discrepancy. Bioinformatics, vol. 22, Issue 14, pp. e49-e57, 2006.
[27] Cheng, Y. C., Chou, T. I., Chiu, S. W., Tang, K. T. A Concentration-based drift calibration transfer learning method for gas sensor array data," IEEE Sensors Letters, vol. 4, no. 10, pp. 1-4, Art no. 7003704, Oct. 2020.
[28] Ziyatdinov, A., Marco, S., Chaudry, A., Persaud, K., Caminal, P., Perera, A. Drift compensation of gas sensor array data by common principal component analysis. Sensors and Actuators B: Chemical, vol. 146, Issue 2, pp. 460-465, 2010.
[29] Saeys, Y., Inza, I., Arrañaga, P. A review of feature selection techniques in bioinformatics. Bioinformatics, vol. 23, no. 19, pp. 2507-2517, 2007.
[30] Gramm, A. S. High performance solvent vapor identification with a two sensor array using temperature cycling and pattern classification. Sensors and Actuators B-Chemical, vol. 95, no. 58, 2003.
[31] Feng, S., Farha, F., Li, Q., Wan, Y., Xu, Y., Zhang, T., Ning, H. Review on Smart Gas Sensing Technology. Sensors 19, no. 17: 3760, 2019.
[32] Zhao, X., Li, P., Xiao, K., Meng, X., Han, L., Yu, C. Sensor Drift Compensation Based on the Improved LSTM and SVM Multi-Class Ensemble Learning Models. Sensors 19, no. 19: 3844, 2019.
[33] Yan, K. and Zhang, D. Calibration transfer and drift compensation of e-noses via
coupled task learning. Sensors and Actuators B: Chemical, vol. 225, pp. 288-297,
2016.
[34] Zhang, L., Tian, F., Liu, S., Dang, L., Peng, X., Yin, X. Chaotic time series prediction of e-nose sensor drift in embedded phase space. Sensors and Actuators B: Chemical, vol. 182, pp. 71–79, 2013.