研究目的
To identify the origin of coal rapidly and efficiently which can effectively strengthen the supervision of coal quality in ports using near-infrared spectroscopy combined with supportive vector machines (SVM) analysis method.
研究成果
Near-infrared spectroscopy combined with the SVM techniques can be used to classify the coal samples according to their origins, with 98.8% with subsequent Linear SVM, Quadratic SVM, Cubic SVM, Fine Gaussian SVM, Medium Gaussian SVM and Coarse Gaussian SVM. This method can be used as a non-destructive tool to validate the origins of coal samples. For this data set investigated, the selected methods (Medium Gaussian SVM) appear to be of particular interest for the classification of coal samples.
1：Experimental Design and Method Selection:
Near infrared spectroscopy combined with supportive vector machines (SVM) analysis method is used to analyze 243 coal samples of different origins. Six supportive vector machines with different kernel functions are employed to discriminate origins of coal samples.
2：Sample Selection and Data Sources:
A total of 243 samples of coals from Australia, Canada, China, Indonesia and Russia are used. The spectra are acquired by Antaris Fourier transform near infrared spectrometer.
3：List of Experimental Equipment and Materials:
Antaris Fourier transform near infrared spectrometer with particle size of coal samples <0.2 mm, 64 scans time, 1557 wavelength points.
4：2 mm, 64 scans time, 1557 wavelength points.
Experimental Procedures and Operational Workflow:
4. Experimental Procedures and Operational Workflow: The spectra of coal samples are acquired and pre-processed using Principal component analysis (PCA). Six supportive vector machines are employed to discriminate origins of coal samples.
5：Data Analysis Methods:
The accuracy for the samples and the training time of SVM is analyzed.
独家科研数据包，助您复现前沿成果，加速创新突破
获取完整内容
