MyMedR

Displaying publications 21 - 40 of 389 in total

Abstract:

Sort:

Fulltext Advanced machine learning model for better prediction accuracy of soil temperature at different depths

Alizamir M, Kisi O, Ahmed AN, Mert C, Fai CM, Kim S, et al.

PLoS One, 2020;15(4):e0231055.
PMID: 32287272 DOI: 10.1371/journal.pone.0231055

Soil temperature has a vital importance in biological, physical and chemical processes of terrestrial ecosystem and its modeling at different depths is very important for land-atmosphere interactions. The study compares four machine learning techniques, extreme learning machine (ELM), artificial neural networks (ANN), classification and regression trees (CART) and group method of data handling (GMDH) in estimating monthly soil temperatures at four different depths. Various combinations of climatic variables are utilized as input to the developed models. The models' outcomes are also compared with multi-linear regression based on Nash-Sutcliffe efficiency, root mean square error, and coefficient of determination statistics. ELM is found to be generally performs better than the other four alternatives in estimating soil temperatures. A decrease in performance of the models is observed by an increase in soil depth. It is found that soil temperatures at three depths (5, 10 and 50 cm) could be mapped utilizing only air temperature data as input while solar radiation and wind speed information are also required for estimating soil temperature at the depth of 100 cm.

Matched MeSH terms: Linear Models
Advancement in biosensor: "Telediagnosis" and "remote digital imaging"

Gopinath SCB, Ismail ZH, Shapiai MI, Yasin MNM

Biotechnol Appl Biochem, 2021 May 19.
PMID: 34009645 DOI: 10.1002/bab.2196

Current developments in sensors and actuators are heralding a new era to facilitate things to happen effortlessly and efficiently with proper communication. On the other hand, Internet of Things (IoT) has been boomed up with er potential and occupies a wide range of disciplines. This study has choreographed to design of an algorithm and a smart data-processing scheme to implement the obtained data from the sensing system to transmit to the receivers. Technically, it is called "telediagnosis" and "remote digital monitoring," a revolution in the field of medicine and artificial intelligence. For the proof of concept, an algorithmic approach has been implemented for telediagnosis with one of the degenerative diseases, that is, Parkinson's disease. Using the data acquired from an improved interdigitated electrode, sensing surface was evaluated with the attained sensitivity of 100 fM (n = 3), and the limit of detection was calculated with the linear regression value coefficient. By the designed algorithm and data processing with the assistance of IoT, further validation was performed and attested the coordination. This proven concept can be ideally used with all sensing strategies for immediate telemedicine by end-to-end communications.

Matched MeSH terms: Linear Models
Age at menarche of university students in Bangladesh: secular trends and association with adult anthropometric measures and socio-demographic factors

Hossain MG, Islam S, Aik S, Zaman TK, Lestrel PE

J Biosoc Sci, 2010 Sep;42(5):677-87.
PMID: 20529410 DOI: 10.1017/S0021932010000210

Age at menarche has been shown to be an important indicator for diseases such as breast cancer and ischaemic heart disease. The aim of the present study was to document secular trends in age at menarche and their association with anthropometric measures and socio-demographic factors in university students in Bangladesh. Data were collected from 995 students from Rajshahi University using a stratified sampling technique between July 2004 and May 2005. Trends in age at menarche were examined by linear regression analysis. Multiple regression analysis was used to assess the association of age at menarche with adult anthropometric measures and various socio-demographic factors. The mean and median age of menarche were 13.12+/-1.16 and 13.17 years, respectively, with an increasing tendency among birth-year cohorts from 1979 to 1986. Menarcheal age was negatively associated with BMI (p<0.01), but positively associated with height (p<0.05). Early menarche was especially pronounced among students from urban environments, Muslims and those with better educated mothers. Increasing age at menarche may be explained by improved nutritional status among Bangladeshi populations. Early menarche was associated with residence location at adolescence, religion and mother's education.

Matched MeSH terms: Linear Models
Age estimation from structural changes of teeth and buccal alveolar bone level

Koh KK, Tan JS, Nambiar P, Ibrahim N, Mutalik S, Khan Asif M

J Forensic Leg Med, 2017 May;48:15-21.
PMID: 28407514 DOI: 10.1016/j.jflm.2017.03.004

Forensic odontology plays a vital role in the identification and age estimation of unknown deceased individuals. The purpose of this study is to estimate the chronological age from Cone-Beam Computed Tomography (CBCT) images by measuring the buccal alveolar bone level (ABL) to the cemento-enamel junction and to investigate the possibility of employing the age-related structural changes of teeth as studied by Gustafson. In addition, this study will determine the forensic reliability of employing CBCT images as a technique for dental age estimation. A total of 284 CBCT images of Malays and Chinese patients (150 females and 134 males), aged from 20 years and above were selected, measured and stages of age-related changes were recorded using the i-CAT Vision software. Lower first premolars of both left and right side of the jaw were chosen and the characteristics described by Gustafson, namely attrition, secondary dentine formation and periodontal recession were evaluated. Linear regression analysis was performed for the buccal bone level and the R values obtained were 0.85 and 0.82 for left and right side respectively. Gustafson's characteristics were analysed using multiple regression analysis with chronological age as the dependent variable. The results of the analysis showed R values ranged from 0.44 to 0.62. Therefore it can be safely concluded that the buccal bone level highly correlated with the chronological age and is consequently the most suitable age-related characteristic for forensic age estimation.

Matched MeSH terms: Linear Models
Age-related rarefaction in retinal vasculature is not linear

Che Azemin MZ, Ab Hamid F, Aminuddin A, Wang JJ, Kawasaki R, Kumar DK

Exp Eye Res, 2013 Nov;116:355-358.
PMID: 24512773 DOI: 10.1016/j.exer.2013.10.010

The fractal dimension is a global measure of complexity and is useful for quantifying anatomical structures, including the retinal vascular network. A previous study found a linear declining trend with aging on the retinal vascular fractal dimension (DF); however, it was limited to the older population (49 years and older). This study aimed to investigate the possible models of the fractal dimension changes from young to old subjects (10-73 years). A total of 215 right-eye retinal samples, including those of 119 (55%) women and 96 (45%) men, were selected. The retinal vessels were segmented using computer-assisted software, and non-vessel fragments were deleted. The fractal dimension was measured based on the log-log plot of the number of grids versus the size. The retinal vascular DF was analyzed to determine changes with increasing age. Finally, the data were fitted to three polynomial models. All three models are statistically significant (Linear: R2 = 0.1270, 213 d.f., p linear regression (p linear and cubic models in a sample with a broader age spectrum.

Matched MeSH terms: Linear Models
Age: the most significant risk for diabetes in Indian populations

Rao PV, Ahuja MM, Trivedi BB, Ramachandran M, Samal KC, Zain AZ, et al.

J Indian Med Assoc, 1998 May;96(5):155-7.
PMID: 9828573
Matched MeSH terms: Linear Models
Fulltext Aggression in adolescents: the biopsychosocial predictors

Fatimah Ahmad Fauzi, Nor Afiah Mohd Zulkefli, Anisah Baharom

Malaysian Journal of Medicine and Health Sciences, 2019;15(104):15-15.
MyJurnal

Introduction: Adolescent aggression has become a serious public health problem with the escalating juvenile cases and violence among secondary school students by inflicting harm to others. The objective of this study was to deter-mine the biopsychosocial predictors of adolescent aggression among Form Four students in Hulu Langat. Methods: Cross-sectional study was conducted by proportionate population sampling method among Form Four students from all public secondary schools in Hulu Langat. Pre-tested questionnaires distributed to measure students’ ag-gression, demographic (ethnicity, family income), biological (sex, head injury, nutritional deficiency, breakfast skip-ping), psychological (attitude towards aggression, normative beliefs to aggression, personality trait, and emotional intelligence), and social factors (family environment, single parent status, domestic violence, peer deviant affiliation, alcohol, smoking, and substance abuse). Data was analysed using multivariate analysis to determine the significant predictors. Results: 480 students from four randomly selected schools participated in the study with response rate of 96.5%. The median aggression score was low, which was 23.00 (IQR=12.00). From simple linear regression, 15 factors had significant relationship with adolescent aggression. The predictors of adolescent aggression were lower family income, Malay ethnicity, nutritional deficiency, attitude towards aggression, and peer deviant affiliation (F [8, 244] =15.980, p-value

Matched MeSH terms: Linear Models
Algorithms that eliminate the effects of calibration artefact and trial-imposed offsets of Masimo oximeter in BOOST-NZ trial

Zahari M, Lee DS, Darlow BA

J Clin Monit Comput, 2016 Oct;30(5):669-78.
PMID: 26282827 DOI: 10.1007/s10877-015-9752-1

The displayed readings of Masimo pulse oximeters used in the Benefits Of Oxygen Saturation Targeting (BOOST) II and related trials in very preterm babies were influenced by trial-imposed offsets and an artefact in the calibration software. A study was undertaken to implement new algorithms that eliminate the effects of offsets and artefact. In the BOOST-New Zealand trial, oxygen saturations were averaged and stored every 10 s up to 36 weeks' post-menstrual age. Two-hundred and fifty-seven of 340 babies enrolled in the trial had at least two weeks of stored data. Oxygen saturation distribution patterns corresponding with a +3 % or -3 % offset in the 85-95 % range were identified together with that due to the calibration artefact. Algorithms involving linear and quadratic interpolations were developed, implemented on each baby of the dataset and validated using the data of a UK preterm baby, as recorded from Masimo oximeters with the original software and a non-offset Siemens oximeter. Saturation distributions obtained were compared for both groups. There were a flat region at saturations 85-87 % and a peak at 96 % from the lower saturation target oximeters, and at 93-95 and 84 % respectively from the higher saturation target oximeters. The algorithms lowered the peaks and redistributed the accumulated frequencies to the flat regions and artefact at 87-90 %. The resulting distributions were very close to those obtained from the Siemens oximeter. The artefact and offsets of the Masimo oximeter's software had been addressed to determine the true saturation readings through the use of novel algorithms. The implementation would enable New Zealand data be included in the meta-analysis of BOOST II trials, and be used in neonatal oxygen studies.

Matched MeSH terms: Linear Models
Altered circulating concentrations of active glucagon-like peptide (GLP-1) and dipeptidyl peptidase 4 (DPP4) in obese subjects and their association with insulin resistance

Ahmed RH, Huri HZ, Muniandy S, Al-Hamodi Z, Al-Absi B, Alsalahi A, et al.

Clin Biochem, 2017 Sep;50(13-14):746-749.
PMID: 28288852 DOI: 10.1016/j.clinbiochem.2017.03.008

OBJECTIVES: Soluble DPP4 (sDPP4) is a novel adipokine that degrades glucagon-like peptide (GLP-1). We evaluated the fasting serum levels of active GLP-1 and sDPP4 in obese, overweight and normal weight subjects to assess the association between sDPP4 levels, active GLP-1 levels and insulin resistance in obese subjects.
METHODS: The study involved 235 Malaysian subjects who were randomly selected (66 normal weight subjects, 97 overweight, 59 obese subjects, and 13 subjects who were underweight). Serum sDPP4 and active GLP-1 levels were examined by enzyme-linked immunosorbent assay (ELISA). Also, body mass index kg/m(2) (BMI), lipid profiles, insulin and glucose levels were evaluated. Insulin resistance (IR) was estimated via the homeostasis model assessment for insulin resistance (HOMA-IR).
RESULTS: Serum sDPP4 levels were significantly higher in obese subjects compared to normal weight subjects (p=0.034), whereas serum levels of active GLP-1 were lower (p=0.021). In obese subjects, sDPP4 levels correlated negatively with active GLP-1 levels (r(2)=-0.326, p=0.015). Furthermore, linear regression showed that sDPP4 levels were positively associated with insulin resistance (B=82.28, p=0.023) in obese subjects.
CONCLUSION: Elevated serum sDPP4 levels and reduced GLP-1 levels were observed in obese subjects. In addition, sDPP4 levels correlated negatively with active GLP-1 levels but was positively associated with insulin resistance. This finding provides evidence that sDPP4 and GLP-1 may play an important role in the pathogenesis of obesity, suggesting that sDPP4 may be valuable as an early marker for the augmented risk of obesity and insulin resistance.

Matched MeSH terms: Linear Models
Fulltext Alzheimer's Disease Determination by a Dual Probe on Gold Nanourchins and Nanohorn Hybrids

Qiu Z, Shen Q, Jiang C, Yao L, Sun X, Li J, et al.

Int J Nanomedicine, 2021;16:2311-2322.
PMID: 33776435 DOI: 10.2147/IJN.S302396

Background: Alzheimer's disease (AD) is a neurodegenerative chronic disorder that causes dementia and problems in thinking, cognitive impairment and behavioral changes. Amyloid-beta (Aβ) is a peptide involved in AD progression, and a high level of Aβ is highly correlated with severe AD. Identifying and quantifying Aβ levels helps in the early treatment of AD and reduces the factors associated with AD.
Materials and Methods: This research introduced a dual probe detection system involving aptamers and antibodies to identify Aβ. Aptamers and antibodies were attached to the gold (Au) urchin and hybrid on the carbon nanohorn-modified surface. The nanohorn was immobilized on the sensor surface by using an amine linker, and then a Au urchin dual probe was immobilized.
Results: This dual probe-modified surface enhanced the current flow during Aβ detection compared with the surface with antibody as the probe. This dual probe interacted with higher numbers of Aβ peptides and reached the detection limit at 10 fM with R2=0.992. Furthermore, control experiments with nonimmune antibodies, complementary aptamer sequences and control proteins did not display the current responses, indicating the specific detection of Aβ.
Conclusion: Aβ-spiked artificial cerebrospinal fluid showed a similar response to current changes, confirming the selective identification of Aβ.

Matched MeSH terms: Linear Models
Fulltext Ambient factors influencing Photovoltaic module temperature in Klang valley

Hedzlin Zainuddin, Maisarah Ismail, Nurul Hidayah Bostamam, Muhamad Mukhzani Muhamad Hanifah, Mohamad Fariz Mohamad Taib, Mohamad Zhafran Hussin

Science Letter, 2016;10(2):23-25.
MyJurnal

The study is conducted to evaluate the significance of solar irradiance, ambient temperature and relative humidity as predictors and to quantify the relative contribution of these ambient parameters as predictors for photovoltaic module temperature model. The module temperature model was developed from experimental data of mono-crystalline and poly-crystalline PV modules retrofitted on metal roof in Klang Valley. The model was developed and analyzed using Multiple Linear Regressions (MLR) and Principle Component Analysis (PCA) Techniques. Solar irradiance, ambient temperature and relative humidity have been proven to be the significant predictors for module temperature. For poly-crystalline PV module, the relative contribution of solar irradiance, ambient temperature and relative humidity are 64.28 %, 17.45 % and 12.64 % respectively. For mono-crystalline PV module, the relative contribution of solar irradiance, ambient temperature and relative humidity are 66.12 %, 17.46 % and 12.48 % respectively. Thus, there is no significant difference in terms of relative contribution of these ambient parameters towards photovoltaic module temperature between poly-crystalline and mono-crystalline PV module technologies.

Matched MeSH terms: Linear Models
An Electrochemical Sandwich Immunosensor for the Detection of HER2 using Antibody-Conjugated PbS Quantum Dot as a label

Lah ZMANH, Ahmad SAA, Zaini MS, Kamarudin MA

J Pharm Biomed Anal, 2019 Sep 10;174:608-617.
PMID: 31265987 DOI: 10.1016/j.jpba.2019.06.024

A facile electrochemical sandwich immunosensor for the detection of a breast cancer biomarker, the human epidermal growth factor receptor 2 (HER2), was designed, using lead sulfide quantum dots-conjugated secondary HER2 antibody (Ab2-PbS QDs) as a label. Using Ab2-PbS QDs in the development of electrochemical immunoassays leads to many advantages such as straightforward synthesis and well-defined stripping signal of Pb(II) through acid dissolution, which in turn yields better sensing performance for the sandwiched immunosensor. In the bioconjugation of PbS QDs, the available amine and hydroxyl groups from secondary anti-HER2 and capped PbS QDs were bound covalently together via carbonyldiimidazole (CDI) acting as a linker. In order to quantify the biomarker, SWV signal was obtained, where the Pb2+ ions after acid dissolution in HCl was detected. The plated mercury film SPCE was also detected in situ. Under optimal conditions, HER2 was detected in a linear range from 1-100 ng/mL with a limit of detection of 0.28 ng/mL. The measures of satisfactory recoveries were 91.3% to 104.3% for the spiked samples, displaying high selectivity. Therefore, this method can be applied to determine HER2 in human serum.

Matched MeSH terms: Linear Models
An ensemble model for forecasting infectious diseases in India

Shashvat K, Basu R, Bhondekar PA, Kaur A

Trop Biomed, 2019 Dec 01;36(4):822-832.
PMID: 33597454

Time series modelling and forecasting plays an important role in various domains. The objective of this paper is to construct a simple average ensemble method to forecast the number of cases for infectious diseases like dengue and typhoid and compare it by applying models for forecasting. In this paper we have also evaluated the correlation between the number of typhoid and dengue cases with the ecological variables. The monthly data of dengue and typhoid cases from 2014 to 2017 were taken from integrated diseases surveillance programme, Government of India. This data was analysed by three models namely support vector regression, neural network and linear regression. The proposed simple average ensemble model was constructed by ensemble of three applied regression models i.e. SVR, NN and LR. We combine the regression models based upon the error metrics such as Mean Square Error, Root Mean Square Error and Mean Absolute Error. It was found that proposed ensemble method performed better in terms of forecast measures. The finding demonstrates that the proposed model outperforms as compared to already available applied models on the basis of forecast accuracy.

Matched MeSH terms: Linear Models
An expectation-maximization algorithm based Kalman smoother approach for single-trial estimation of event-related potentials

Ting CM, Samdin SB, Salleh ShH, Omar MH, Kamarulafizam I

Annu Int Conf IEEE Eng Med Biol Soc, 2012;2012:6534-8.
PMID: 23367426 DOI: 10.1109/EMBC.2012.6347491

This paper applies an expectation-maximization (EM) based Kalman smoother (KS) approach for single-trial event-related potential (ERP) estimation. Existing studies assume a Markov diffusion process for the dynamics of ERP parameters which is recursively estimated by optimal filtering approaches such as Kalman filter (KF). However, these studies only consider estimation of ERP state parameters while the model parameters are pre-specified using manual tuning, which is time-consuming for practical usage besides giving suboptimal estimates. We extend the KF approach by adding EM based maximum likelihood estimation of the model parameters to obtain more accurate ERP estimates automatically. We also introduce different model variants by allowing flexibility in the covariance structure of model noises. Optimal model selection is performed based on Akaike Information Criterion (AIC). The method is applied to estimation of chirp-evoked auditory brainstem responses (ABRs) for detection of wave V critical for assessment of hearing loss. Results shows that use of more complex covariances are better estimating inter-trial variability.

Matched MeSH terms: Linear Models
An offset-free MPC formulation for nonlinear systems using adaptive integral controller

Hermansson AW, Syafiie S

ISA Trans, 2019 Aug;91:66-77.
PMID: 30782432 DOI: 10.1016/j.isatra.2019.01.037

This paper investigates a novel offset-free control scheme based on a multiple model predictive controller (MMPC) and an adaptive integral action controller for nonlinear processes. Firstly, the multiple model description captures the essence of the nonlinear process, while keeping the MPC optimization linear. Multiple models also enable the controller to deal with the uncertainty associated with changing setpoint. Then, a min-max approach is utilized to counter the effect of parametric uncertainty between the linear models and the nonlinear process. Finally, to deal with other uncertainties, such as input and output disturbances, an adaptive integral action controller is run in parallel to the MMPC. Thus creating a novel offset-free approach for nonlinear systems that is more easily tuned than observer-based MPC. Simulation results for a pH-controller, which acts as an example of a nonlinear process, are presented to demonstrate the usefulness of the technique compared to using an observer-based MPC.

Matched MeSH terms: Linear Models
Analysis of infant cry through weighted linear prediction cepstral coefficients and Probabilistic Neural Network

Hariharan M, Chee LS, Yaacob S

J Med Syst, 2012 Jun;36(3):1309-15.
PMID: 20844933 DOI: 10.1007/s10916-010-9591-z

Acoustic analysis of infant cry signals has been proven to be an excellent tool in the area of automatic detection of pathological status of an infant. This paper investigates the application of parameter weighting for linear prediction cepstral coefficients (LPCCs) to provide the robust representation of infant cry signals. Three classes of infant cry signals were considered such as normal cry signals, cry signals from deaf babies and babies with asphyxia. A Probabilistic Neural Network (PNN) is suggested to classify the infant cry signals into normal and pathological cries. PNN is trained with different spread factor or smoothing parameter to obtain better classification accuracy. The experimental results demonstrate that the suggested features and classification algorithms give very promising classification accuracy of above 98% and it expounds that the suggested method can be used to help medical professionals for diagnosing pathological status of an infant from cry signals.

Matched MeSH terms: Linear Models
Analyzing the relationships between travel mode indicators and the number of passenger transport fatalities at the city level

Asadi-Shekari Z, Moeinaddini M, Sultan Z, Shah MZ, Hamzah A

Traffic Inj Prev, 2016 08 17;17(6):650-5.
PMID: 26890058 DOI: 10.1080/15389588.2015.1136739

OBJECTIVE: A number of efforts have been conducted on travel behavior and transport fatalities at the neighborhood or street level, and they have identified different factors such as roadway characteristics, personal indicators, and design indicators related to transport safety. However, only a limited number of studies have considered the relationship between travel behavior indicators and the number of transport fatalities at the city level. Therefore, this study explores this relationship and how to fill the mentioned gap in current knowledge.
METHOD: A generalized linear model (GLM) estimates the relationships between different travel mode indicators (e.g., length of motorway per inhabitants, number of motorcycles per inhabitant, percentage of daily trips on foot and by bicycle, percentage of daily trips by public transport) and the number of passenger transport fatalities. Because this city-level model is developed using data sets from different cities all over the world, the impacts of gross domestic product (GDP) are also included in the model.
CONCLUSIONS: Overall, the results imply that the percentage of daily trips by public transport, the percentage of daily trips on foot and by bicycle, and the GDP per inhabitant have negative relationships with the number of passenger transport fatalities, whereas motorway length and the number of motorcycles have positive relationships with the number of passenger transport fatalities.

Matched MeSH terms: Linear Models
Fulltext Angular velocity and elevation angle: the proposed human model scalable tracking model using linear regression

Yong CY, Sudirman R, Chew KM

Sains Malaysiana, 2015;44(12):1661-1669.

A scalable tracking human model was proposed for recognizing human jogging and walking activities. The model aims to detect and track a particular subject by using wearable sensor. Data collected are in accelerometer readings in three axes and gyroscope readings in three axes. The development of proposed human model is based on the moderating effects on human movements. Two moderators were proposed as the moderating factors of human motion and they are angular velocity and elevation angle. Linear regression is used to investigate the relationship among inputs, moderators and outputs of the model. The result of this study showed that the angular velocity and elevation angle moderators are affecting the relation of research output. Acceleration in x-axis (Ax) and angular velocity in y-axis (Gy) are the two main components in directing
a motion. Classification between jogging and walking motions was done by measuring the magnitude of angular velocity and elevation angle. Jogging motion was classified and identified with larger angular velocity and elevation angle. The two proposed hypotheses were supported and proved by research output. The result is expected to be beneficial and able to assist researcher in investigating human motions.

Matched MeSH terms: Linear Models
Fulltext Anthropometric and cardiometabolic risk factors in parents and child obesity in Segamat, Malaysia

Partap U, Young EH, Allotey P, Sandhu MS, Reidpath DD

Int J Epidemiol, 2017 Oct 01;46(5):1523-1532.
PMID: 29106558 DOI: 10.1093/ije/dyx114

BACKGROUND: There is little evidence regarding risk factors for child obesity in Asian populations, including the role of parental anthropometric and cardiometabolic risk factors. We examined the relation between parental risk factors and child obesity in a Malaysian population.
METHODS: We used data from health and demographic surveillance conducted by the South East Asia Community Observatory in Segamat, Malaysia. Analyses included 9207 individuals (4806 children, 2570 mothers and 1831 fathers). Child obesity was defined based on the World Health Organization 2007 reference. We assessed the relation between parental anthropometric (overweight, obesity and central obesity) and cardiometabolic (systolic hypertension, diastolic hypertension and hyperglycaemia) risk factors and child obesity, using mixed effects Poisson regression models with robust standard errors.
RESULTS: We found a high burden of overweight and obesity among children in this population (30% overweight or obese). Children of one or more obese parents had a 2-fold greater risk of being obese compared with children of non-obese parents. Sequential adjustment for parental and child characteristics did not materially affect estimates (fully adjusted relative risk for obesity in both parents: 2.39, 95% confidence interval: 1.82, 3.10, P

Matched MeSH terms: Linear Models
Fulltext Application of Neural Network in Predicting H2S from an Acid Gas Removal Unit (AGRU) with Different Compositions of Solvents

Hakimi M, Omar MB, Ibrahim R

Sensors (Basel), 2023 Jan 16;23(2).
PMID: 36679816 DOI: 10.3390/s23021020

The gas sweetening process removes hydrogen sulfide (H2S) in an acid gas removal unit (AGRU) to meet the gas sales' specification, known as sweet gas. Monitoring the concentration of H2S in sweet gas is crucial to avoid operational and environmental issues. This study shows the capability of artificial neural networks (ANN) to predict the concentration of H2S in sweet gas. The concentration of N-methyldiethanolamine (MDEA) and Piperazine (PZ), temperature and pressure as inputs, and the concentration of H2S in sweet gas as outputs have been used to create the ANN network. Two distinct backpropagation techniques with various transfer functions and numbers of neurons were used to train the ANN models. Multiple linear regression (MLR) was used to compare the outcomes of the ANN models. The models' performance was assessed using the mean absolute error (MAE), root mean square error (RMSE), and coefficient of determination (R2). The findings demonstrate that ANN trained by the Levenberg-Marquardt technique, equipped with a logistic sigmoid (logsig) transfer function with three neurons achieved the highest R2 (0.966) and the lowest MAE (0.066) and RMSE (0.122) values. The findings suggested that ANN can be a reliable and accurate prediction method in predicting the concentration of H2S in sweet gas.

Matched MeSH terms: Linear Models

Filters

Please provide feedback to Administrator (afdal@afpm.org.my)

External Links