Unlock instant, AI-driven research and patent intelligence for your innovation.

Bagging extreme learning machine ensemble modeling method

A technology of extreme learning machine and modeling method, applied in the field of chemometrics, can solve problems such as poor stability of ELM, and achieve the effect of improving prediction accuracy and stability

Active Publication Date: 2018-05-15
TIANJIN POLYTECHNIC UNIV +1
View PDF5 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0004] Combining the advantages of ELM and Bagging, the present invention proposes a Bagging-based ELM integrated modeling method, which is used for quantitative analysis of complex samples, which not only retains the advantages of fast calculation speed and strong predictive ability of ELM, but also overcomes the disadvantages of poor stability of ELM

Method used

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View more

Image

Smart Image Click on the blue labels to locate them in the text.
Viewing Examples
Smart Image
  • Bagging extreme learning machine ensemble modeling method
  • Bagging extreme learning machine ensemble modeling method
  • Bagging extreme learning machine ensemble modeling method

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0039] This embodiment is applied to ultraviolet spectrum analysis to determine the content of monoaromatic compounds in a fuel sample. The specific steps are as follows:

[0040] (1) The ultraviolet spectrum data of 115 fuel samples were collected, with a wavelength range of 200-400nm and a sampling interval of 0.35nm, including 572 wavelength points. The spectra were measured by a Varian Cary 3 UV-visible spectrometer. The content of monoaromatic compounds was determined by HPG1205A supercritical fluid chromatography, carbon dioxide was used as carrier gas, and the flow rate was 2mLmin -1 , the furnace temperature is 35ºC, the outlet pressure is 150bar, and the detector is a flame ionization detector. According to the division of the data set on the website, 70 samples are used as the training set and 45 samples are used as the prediction set.

[0041] (2) Perform boostrap resampling on the training set samples, and randomly select a certain number of samples as a training...

Embodiment 2

[0050] This embodiment is applied to ultraviolet spectrum analysis to determine the content value of ethanol components. The specific steps are as follows:

[0051] (1) Collect near-infrared spectral data of 95 ethanol solution samples, the wavelength range is 850-1049nm, the sampling interval is 1nm, including 200 wavelength points, and the spectrum is measured by HP 8453 spectrometer. According to the division of the data set on the website, 65 samples are used as the training set and 30 samples are used as the prediction set.

[0052] (2) Perform boostrap resampling on the training set samples, and randomly select a certain number of samples as a training subset.

[0053] (3) Determine the optimal excitation function of the extreme learning machine and the number of hidden layer nodes, and use the samples of the training subset to establish the extreme learning machine sub-model.

[0054] Repeat steps (2)-(3) multiple times to create multiple sub-models.

[0055] (4) For...

Embodiment 3

[0061] This embodiment is applied to near-infrared spectrum analysis to measure the density value of a diesel sample. The specific steps are as follows:

[0062] (1) Collect near-infrared spectral data of 263 diesel engine fuel samples, the wavelength range is 750-1550nm, the sampling interval is 2nm, including 401 wavelength points, the data is provided by the US military Southwest Research Institute (SWRI), San Antonio, TX through Eigenvector Research, Inc. (Manson, Washington, available for download at http: / / www.eigenvector.com / Data / SWRI). According to the division of the dataset on the website, 142 samples are used as the training set and 121 samples are used as the prediction set.

[0063] (2) Perform boostrap resampling on the training set samples, and randomly select a certain number of samples as a training subset.

[0064] (3) Determine the optimal excitation function of the extreme learning machine and the number of hidden layer nodes, and use the samples of the t...

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More

PUM

No PUM Login to View More

Abstract

The invention belongs to the technical field of chemometrics, in particular to a Bagging extreme learning machine integrated modeling method. The specific steps of the present invention are: collecting the sample spectral data of the measured object, determining the content of the measured component of the sample; dividing the sample set into a training set and a prediction set; performing boostrap resampling on the training set samples, and randomly selecting a certain number of samples as a Training subsets; use the samples of the training subsets to establish an extreme learning machine sub-model; repeat multiple times to establish multiple sub-models; for unknown samples, simply average the prediction results of multiple sub-models to obtain the final prediction result. Compared with the ELM method, the method of the present invention has obvious advantages in prediction accuracy and stability. The invention is applicable to the field of quantitative analysis of complex substances such as petroleum, tobacco, food, and traditional Chinese medicine.

Description

technical field [0001] The invention belongs to the technical field of chemometrics, and in particular relates to a Bagging extreme learning machine integrated modeling method. Background technique [0002] Artificial neural networks have been widely used in various fields such as biology, chemistry, medicine, and economy because of their powerful self-adaptation, self-organization, self-learning, and nonlinear mapping capabilities. However, traditional neural network learning algorithms (such as BP algorithm) need to manually set a large number of network training parameters, the training speed is slow, and it is easy to generate local optimal solutions. In 2004, Professor Huang Guangbin of Nanyang Technological University in Singapore proposed a new algorithm for a single hidden layer feed-forward neural network, named Extreme Learning Machine (ELM). The core of the ELM algorithm is to change the training problem of the neural network into the problem of solving the least...

Claims

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More

Application Information

Patent Timeline
no application Login to View More
Patent Type & Authority Patents(China)
IPC IPC(8): G06F17/50G06N3/08
Inventor 卞希慧李淑娟谭小耀王江江王治国刘维国陈宗蓬王晨
Owner TIANJIN POLYTECHNIC UNIV