Limits...
Pattern classification using an olfactory model with PCA feature selection in electronic noses: study and application.

Fu J, Huang C, Xing J, Zheng J - Sensors (Basel) (2012)

Bottom Line: The principal component analysis technique was applied for feature selection and dimension reduction.In the former case the results showed that the average correct classification rate increased as more principal components were put in to feature vector.We concluded that 6~8 channels of the model with principal component feature vector values of at least 90% cumulative variance is adequate for a classification task of 3~5 pattern classes considering the trade-off between time consumption and classification rate.

View Article: PubMed Central - PubMed

Affiliation: College of Computer Science & Information Engineering, Zhejiang Gongshang University, Hangzhou, China. junfu@zjgsu.edu.cn

ABSTRACT
Biologically-inspired models and algorithms are considered as promising sensor array signal processing methods for electronic noses. Feature selection is one of the most important issues for developing robust pattern recognition models in machine learning. This paper describes an investigation into the classification performance of a bionic olfactory model with the increase of the dimensions of input feature vector (outer factor) as well as its parallel channels (inner factor). The principal component analysis technique was applied for feature selection and dimension reduction. Two data sets of three classes of wine derived from different cultivars and five classes of green tea derived from five different provinces of China were used for experiments. In the former case the results showed that the average correct classification rate increased as more principal components were put in to feature vector. In the latter case the results showed that sufficient parallel channels should be reserved in the model to avoid pattern space crowding. We concluded that 6~8 channels of the model with principal component feature vector values of at least 90% cumulative variance is adequate for a classification task of 3~5 pattern classes considering the trade-off between time consumption and classification rate.

Show MeSH

Related in: MedlinePlus

The average correct classification rate with respect to cumulative variance percentage of principal components.
© Copyright Policy
Related In: Results  -  Collection

License
getmorefigures.php?uid=PMC3376605&req=5

f7-sensors-12-02818: The average correct classification rate with respect to cumulative variance percentage of principal components.

Mentions: Five samples in each class of wine data set were randomly chosen for training set and the others were used for testing. The first n principal components are selected as input feature vector for n-channel KIII model. The parameters of the KIII model in this paper are hhab = 0.5487, hHeb = 0.0395, RHeb = 1.5 and the others are from [27]. The 4th-order Runge-Kutta method with a fixed step of one (1,200 steps total) was applied for numerical integration of the ODEs. The average correct classification rates of ten trials are presented in Table 1, as well as in Figure 6. From the plot, it is very clear that the classification performance of the KIII model gradually become better and better as PC numbers increasing. Figure 7 shows the average correct classification rates of three classes with the increasing of cumulative variance percentage (CVP). Roughly speaking, more than 85% of CVP would represent sufficient information of the data set, and could get 90% of correct classification rate. However, the running time of KIII model implementation increased correspondingly. Considering the trade-off between time consumption and classification rate, the optimal dimension of feature vector is about seven in this classification task.


Pattern classification using an olfactory model with PCA feature selection in electronic noses: study and application.

Fu J, Huang C, Xing J, Zheng J - Sensors (Basel) (2012)

The average correct classification rate with respect to cumulative variance percentage of principal components.
© Copyright Policy
Related In: Results  -  Collection

License
Show All Figures
getmorefigures.php?uid=PMC3376605&req=5

f7-sensors-12-02818: The average correct classification rate with respect to cumulative variance percentage of principal components.
Mentions: Five samples in each class of wine data set were randomly chosen for training set and the others were used for testing. The first n principal components are selected as input feature vector for n-channel KIII model. The parameters of the KIII model in this paper are hhab = 0.5487, hHeb = 0.0395, RHeb = 1.5 and the others are from [27]. The 4th-order Runge-Kutta method with a fixed step of one (1,200 steps total) was applied for numerical integration of the ODEs. The average correct classification rates of ten trials are presented in Table 1, as well as in Figure 6. From the plot, it is very clear that the classification performance of the KIII model gradually become better and better as PC numbers increasing. Figure 7 shows the average correct classification rates of three classes with the increasing of cumulative variance percentage (CVP). Roughly speaking, more than 85% of CVP would represent sufficient information of the data set, and could get 90% of correct classification rate. However, the running time of KIII model implementation increased correspondingly. Considering the trade-off between time consumption and classification rate, the optimal dimension of feature vector is about seven in this classification task.

Bottom Line: The principal component analysis technique was applied for feature selection and dimension reduction.In the former case the results showed that the average correct classification rate increased as more principal components were put in to feature vector.We concluded that 6~8 channels of the model with principal component feature vector values of at least 90% cumulative variance is adequate for a classification task of 3~5 pattern classes considering the trade-off between time consumption and classification rate.

View Article: PubMed Central - PubMed

Affiliation: College of Computer Science & Information Engineering, Zhejiang Gongshang University, Hangzhou, China. junfu@zjgsu.edu.cn

ABSTRACT
Biologically-inspired models and algorithms are considered as promising sensor array signal processing methods for electronic noses. Feature selection is one of the most important issues for developing robust pattern recognition models in machine learning. This paper describes an investigation into the classification performance of a bionic olfactory model with the increase of the dimensions of input feature vector (outer factor) as well as its parallel channels (inner factor). The principal component analysis technique was applied for feature selection and dimension reduction. Two data sets of three classes of wine derived from different cultivars and five classes of green tea derived from five different provinces of China were used for experiments. In the former case the results showed that the average correct classification rate increased as more principal components were put in to feature vector. In the latter case the results showed that sufficient parallel channels should be reserved in the model to avoid pattern space crowding. We concluded that 6~8 channels of the model with principal component feature vector values of at least 90% cumulative variance is adequate for a classification task of 3~5 pattern classes considering the trade-off between time consumption and classification rate.

Show MeSH
Related in: MedlinePlus