Optimal Sensor Selection for Classifying a Set of Ginsengs Using Metal-Oxide Sensors.
Bottom Line:
The relation of the minimum numbers of sensors with number of samples in the sample set was revealed.The results showed that as the number of samples increased, the average minimum number of sensors increased, while the increment decreased gradually and the average optimal classification rate decreased gradually.Moreover, a new approach of sensor selection was proposed to estimate and compare the effective information capacity of each sensor.
View Article:
PubMed Central - PubMed
Affiliation: State Key Laboratory of Industrial Control Technology, Institute of Cyber Systems and Control, Zhejiang University, Hangzhou 310027, Zhejiang, China. jiacheng@zju.edu.cn.
ABSTRACT
The sensor selection problem was investigated for the application of classification of a set of ginsengs using a metal-oxide sensor-based homemade electronic nose with linear discriminant analysis. Samples (315) were measured for nine kinds of ginsengs using 12 sensors. We investigated the classification performances of combinations of 12 sensors for the overall discrimination of combinations of nine ginsengs. The minimum numbers of sensors for discriminating each sample set to obtain an optimal classification performance were defined. The relation of the minimum numbers of sensors with number of samples in the sample set was revealed. The results showed that as the number of samples increased, the average minimum number of sensors increased, while the increment decreased gradually and the average optimal classification rate decreased gradually. Moreover, a new approach of sensor selection was proposed to estimate and compare the effective information capacity of each sensor. No MeSH data available. |
Related In:
Results -
Collection
License getmorefigures.php?uid=PMC4541866&req=5
Mentions: The minimum numbers of sensor Nmin (A) and corresponding average classification performances for all potential sample sets were obtained with the method mentioned in Section 3.2 and shown in Figure 6a. For different sample sets with the same number of samples, Nmin (A) varies a lot, since the complexity for discriminating different sample sets is different. By setting different sample sets with different numbers of sample species, we wanted to investigate how the minimum number of sensors varies as the number of sample species increases. When we averaged Nmin (A) and the corresponding classification performances, the result is shown in Figure 6b. It is observed that as the number of samples (M) increases,average minimum number of sensors Nave (M) also increases, whereas the average classification accuracy decreases smoothly. For example, when M = 2, Nave (2) is 2.9, and corresponding average optimal classification rate is 99.0%, and when M = 3, Nave (3) is 4.5, and corresponding average optimal classification rate is 99.1%. We also noted that the increment of Nave (M) decreases as M increases. The increment of Nave (M) is about 1.6 when M grows from 2 to 3, whereas it is just about 0.4 when M grows for 8 to 9. We deduced that when we gradually add more ginseng samples to the existing nine ginsengs, Nave (M) will gradually to be approximately constant if Nave (M) doesn’t reach 12, and the average optimal classification performance of sample sets with M samples will still decrease gradually. Because when M increases, the classification problem will become more complex, but all existing sensors would not provide sufficient discriminant information, then Nave (M) gradually will become approximately constant while the classification performance gradually decreases. Furthermore, we deduced that if there were a large set of samples and sufficient but not excessive sensors, when the number of samples to discriminate increased, the minimum number of sensors would increase and gradually verge to constant while the optimal classification performance decreased gradually. |
View Article: PubMed Central - PubMed
Affiliation: State Key Laboratory of Industrial Control Technology, Institute of Cyber Systems and Control, Zhejiang University, Hangzhou 310027, Zhejiang, China. jiacheng@zju.edu.cn.
No MeSH data available.