Javascript must be enabled to continue!
FEATURE EXTRACTION AND SUPERVISED LEARNING FOR VOLATILE ORGANIC COMPOUNDS GAS RECOGNITION
View through CrossRef
The emergence of advanced technologies, particularly in the field of artificial intelligence (AI), has sparked significant interest in exploring their potential benefits for various industries, including healthcare. In the medical sector, the utilization of sensing systems has proven valuable for diagnosing pulmonary diseases by detecting volatile organic compounds (VOCs) in exhaled breath. However, the identification of the most informative and discriminating features from VOC sensor arrays remains an unresolved challenge, essential for achieving robust VOC class recognition. This research project aims to investigate effective feature extraction techniques that can be employed as discriminative features for machine learning algorithms. A preliminary dataset was used to predict VOC classification through the application of five supervised machine learning algorithms: k-Nearest Neighbors (kNN), Random Forest (RF), Support Vector Machines (SVM), Logistic Regression (LR), and Artificial Neural Networks (ANN). Ten feature extraction methods were proposed based on changes in sensor response as inputs to classify three types of gases in the dataset. The performance of each model was evaluated and compared using k-Fold cross-validation (k=10) and metrics derived from the confusion matrix. The results demonstrate that the RF model achieved the highest mean accuracy and standard deviation, with values of 0.813 ± 0.035, followed closely by kNN with 0.803 ± 0.033. Conversely, LR, SVM (kernel=Polynomial), and ANN exhibited poor performances when applied to the VOC dataset, with accuracies of 0.447 ± 0.035, 0.403 ± 0.041, and 0.419 ± 0.035, respectively. Therefore, this paper provides evidence that classifying VOC gases based on sensor responses is feasible and emphasizes the need for further research to explore sensor array analysis to enhance feature extraction techniques.
ABSTRAK: Perkembangan teknologi canggih, khususnya dalam bidang kecerdasan buatan (AI), telah mencetuskan minat yang ketara dalam menerokai manfaatnya untuk pelbagai industri, termasuk bidang kesihatan. Dalam sektor perubatan, penggunaan sistem penderiaan telah terbukti bernilai untuk mendiagnosis penyakit paru-paru dengan mengesan sebatian organik meruap (VOC) dalam nafas yang dihembus manusia. Walau bagaimanapun, pengenalpastian ciri yang paling bermaklumat dan mendiskriminasi daripada penderia VOC kekal sebagai cabaran yang tidak dapat diselesaikan, penting untuk mencapai pengiktirafan kelas VOC yang kukuh. Projek penyelidikan ini bertujuan untuk menyiasat teknik pengekstrakan ciri yang berkesan yang boleh digunakan sebagai ciri diskriminatif untuk algoritma pembelajaran mesin. Set data awal digunakan untuk meramalkan klasifikasi VOC melalui aplikasi lima algoritma pembelajaran mesin yang diselia: k-Nearest Neighbors (kNN), Random Forest (RF), Support Vector Machines (SVM), Logistic Regression (LR), dan Artificial Neural Networks (ANN). Sepuluh kaedah pengekstrakan ciri telah dicadangkan berdasarkan perubahan dalam tindak balas penderia sebagai input untuk mengklasifikasikan tiga jenis gas dalam set data. Prestasi setiap model telah dinilai dan dibandingkan menggunakan pengesahan silang k-Fold (k=10) dan metrik yang diperoleh daripada confusion matriks . Keputusan menunjukkan bahawa model RF mencapai ketepatan minima tertinggi dan sisihan piawai, dengan nilai 0.813 ± 0.035, diikuti oleh kNN dengan 0.803 ± 0.033. Sebaliknya, LR, SVM (kernel=Polinomial), dan ANN mempamerkan prestasi yang lemah apabila digunakan pada dataset VOC, dengan ketepatan masing-masing 0.447 ± 0.035, 0.403 ± 0.041 dan 0.419 ± 0.035. Oleh itu, kertas kerja ini memberikan bukti bahawa mengklasifikasikan gas VOC berdasarkan tindak balas penderia adalah boleh dilaksanakan dan menekankan keperluan untuk penyelidikan lanjut untuk meneroka analisis tatasusunan penderia untuk meningkatkan teknik pengekstrakan ciri.
Title: FEATURE EXTRACTION AND SUPERVISED LEARNING FOR VOLATILE ORGANIC COMPOUNDS GAS RECOGNITION
Description:
The emergence of advanced technologies, particularly in the field of artificial intelligence (AI), has sparked significant interest in exploring their potential benefits for various industries, including healthcare.
In the medical sector, the utilization of sensing systems has proven valuable for diagnosing pulmonary diseases by detecting volatile organic compounds (VOCs) in exhaled breath.
However, the identification of the most informative and discriminating features from VOC sensor arrays remains an unresolved challenge, essential for achieving robust VOC class recognition.
This research project aims to investigate effective feature extraction techniques that can be employed as discriminative features for machine learning algorithms.
A preliminary dataset was used to predict VOC classification through the application of five supervised machine learning algorithms: k-Nearest Neighbors (kNN), Random Forest (RF), Support Vector Machines (SVM), Logistic Regression (LR), and Artificial Neural Networks (ANN).
Ten feature extraction methods were proposed based on changes in sensor response as inputs to classify three types of gases in the dataset.
The performance of each model was evaluated and compared using k-Fold cross-validation (k=10) and metrics derived from the confusion matrix.
The results demonstrate that the RF model achieved the highest mean accuracy and standard deviation, with values of 0.
813 ± 0.
035, followed closely by kNN with 0.
803 ± 0.
033.
Conversely, LR, SVM (kernel=Polynomial), and ANN exhibited poor performances when applied to the VOC dataset, with accuracies of 0.
447 ± 0.
035, 0.
403 ± 0.
041, and 0.
419 ± 0.
035, respectively.
Therefore, this paper provides evidence that classifying VOC gases based on sensor responses is feasible and emphasizes the need for further research to explore sensor array analysis to enhance feature extraction techniques.
ABSTRAK: Perkembangan teknologi canggih, khususnya dalam bidang kecerdasan buatan (AI), telah mencetuskan minat yang ketara dalam menerokai manfaatnya untuk pelbagai industri, termasuk bidang kesihatan.
Dalam sektor perubatan, penggunaan sistem penderiaan telah terbukti bernilai untuk mendiagnosis penyakit paru-paru dengan mengesan sebatian organik meruap (VOC) dalam nafas yang dihembus manusia.
Walau bagaimanapun, pengenalpastian ciri yang paling bermaklumat dan mendiskriminasi daripada penderia VOC kekal sebagai cabaran yang tidak dapat diselesaikan, penting untuk mencapai pengiktirafan kelas VOC yang kukuh.
Projek penyelidikan ini bertujuan untuk menyiasat teknik pengekstrakan ciri yang berkesan yang boleh digunakan sebagai ciri diskriminatif untuk algoritma pembelajaran mesin.
Set data awal digunakan untuk meramalkan klasifikasi VOC melalui aplikasi lima algoritma pembelajaran mesin yang diselia: k-Nearest Neighbors (kNN), Random Forest (RF), Support Vector Machines (SVM), Logistic Regression (LR), dan Artificial Neural Networks (ANN).
Sepuluh kaedah pengekstrakan ciri telah dicadangkan berdasarkan perubahan dalam tindak balas penderia sebagai input untuk mengklasifikasikan tiga jenis gas dalam set data.
Prestasi setiap model telah dinilai dan dibandingkan menggunakan pengesahan silang k-Fold (k=10) dan metrik yang diperoleh daripada confusion matriks .
Keputusan menunjukkan bahawa model RF mencapai ketepatan minima tertinggi dan sisihan piawai, dengan nilai 0.
813 ± 0.
035, diikuti oleh kNN dengan 0.
803 ± 0.
033.
Sebaliknya, LR, SVM (kernel=Polinomial), dan ANN mempamerkan prestasi yang lemah apabila digunakan pada dataset VOC, dengan ketepatan masing-masing 0.
447 ± 0.
035, 0.
403 ± 0.
041 dan 0.
419 ± 0.
035.
Oleh itu, kertas kerja ini memberikan bukti bahawa mengklasifikasikan gas VOC berdasarkan tindak balas penderia adalah boleh dilaksanakan dan menekankan keperluan untuk penyelidikan lanjut untuk meneroka analisis tatasusunan penderia untuk meningkatkan teknik pengekstrakan ciri.
Related Results
Efficient and Effective Gas Sensor Calibration with Randomized Gas Mixtures
Efficient and Effective Gas Sensor Calibration with Randomized Gas Mixtures
Introduction
The selective quantification of target gases in complex mixtures is an important part of numerous applications of chemical gas sensors. ...
Unconventional Reservoirs: Basic Petrophysical Concepts for Shale Gas
Unconventional Reservoirs: Basic Petrophysical Concepts for Shale Gas
Abstract
Unconventional reservoirs have burst with considerable force in oil and gas production worldwide. Shale Gas is one of them, with intense activity taking pla...
Comparisons of Pore Structure for Unconventional Tight Gas, Coalbed Methane and Shale Gas Reservoirs
Comparisons of Pore Structure for Unconventional Tight Gas, Coalbed Methane and Shale Gas Reservoirs
Extended abstract
Tight sands gas, coalbed methane and shale gas are three kinds of typical unconventional natural gas. With the decrease of conventional oil and gas...
Characterization of berry and wine aroma development in interspecific hybrids important to Missouri using a metabolomics based approach
Characterization of berry and wine aroma development in interspecific hybrids important to Missouri using a metabolomics based approach
With increases in climate extremes resulting in more abiotic and biotic stress on crops such as grapes, it is essential to develop new cultivars that are more robust than the tradi...
Critical Gas Saturation During Depressurisation and its Importance in the Brent Field
Critical Gas Saturation During Depressurisation and its Importance in the Brent Field
Critical Gas Saturation During Depressurisation and its Importance in the Brent Field.
Abstract
After some 20 years of pressure ...
Capturing actively produced microbial volatile organic compounds from human associated samples with vacuum assisted sorbent extraction
Capturing actively produced microbial volatile organic compounds from human associated samples with vacuum assisted sorbent extraction
ABSTRACTVolatile organic compounds (VOC) from biological samples have unknown origins. VOCs may originate from the host or different microbial organisms from a microbial community....
Understanding Unconventional Gas Reservoir Damages
Understanding Unconventional Gas Reservoir Damages
Abstract
It is estimated that there are large reserves of unconventional gas located throughout the world, including coalbed methane, shale gas and tight gas sand...
Effects of Coal Permeability Anisotropy on Gas Extraction Performance
Effects of Coal Permeability Anisotropy on Gas Extraction Performance
To investigate gas flow characteristics in coal seams with strong anisotropy, a coupled anisotropic dual-porosity model was established. Effects of permeability anisotropy on varia...

