Azian Azamimi Abdullah
Thumbnail Image
Preferred name
Azian Azamimi Abdullah
Official Name
Azian Azamimi, Abdullah
Alternative Name
Abdullah, Azian Azamimi
Abdullah, A. A.
Main Affiliation
Scopus Author ID
36968589600
Researcher ID
F-5186-2010

Research Output

Filters

11
Reset filters

Settings
Now showing 1 - 10 of 11
Thumbnail Image
Publication

Performance Comparison of Machine Learning Algorithms for Classification of Chronic Kidney Disease (CKD)

2020-06-17 , Azian Azamimi Abdullah , Hafidz S.A. , Wan Khairunizam Wan Ahmad

Kidney is one of the vital organs in a human body while ironically, chronic kidney disease (CKD) is one of the main causes of death in the world. Due to the low rate of loss of kidney function, the disease is often overlooked until it is in a really bad condition. Dysfunctional kidney may lead to accumulation of wastes in blood which would affect several other systems and functions of the body such as blood pressure, red blood cell production, vitamin D and bone health. Machine learning algorithms can help in classifying the patients who have CKD or not. Even though several studies have been made to classify CKD on patients using machine-learning tool, not many researchers perform pre-processing and feature selection technique to obtain quality and dependable result. Machine learning used with feature selection techniques are shown to have better and more dependable result. In this study, feature selection methods such as Random Forest feature selection, forward selection, forward exhaustive selection, backward selection and backward exhaustive selection were identified and evaluated. Then, machine learning classifiers such as Random Forest, Linear and Radial SVM, Naïve Bayes and Logistic Regression were implemented. Lastly, the performance of each machine-learning model was evaluated in terms of accuracy, sensitivity, specificity and AUC score. The results showed that Random Forest classifier with Random Forest feature selection is the most suitable machine learning model for classification of CKD as it has the highest accuracy, sensitivity, specificity and AUC with 98.825%, 98.04%, 100% and 98.9% respectively which outperformed other classifiers.

View
No Thumbnail Available
Publication

In silico prediction of biological activity of volatile metabolite using deep learning algorithm

2023 , Azian Azamimi Abdullah , Nur Najwa Amrahuddin , Shigehiko Kanaya

Recently, researchers have shown an increased interest in studying volatile metabolite as disease biomarker which can provide non-invasive, safer and cheaper diagnosis tools. Volatile metabolites are also being produced during secondary metabolite and they are not directly involved in normal growth or reproduction of cell but crucial for the survival in the environment. These metabolites play vital role in defensing plants from herbivores, attracting pollinators, seed dispersers and also plant-plant interaction. It is important to investigate the role of volatile metabolites in human healthcare and chemical ecology, however a limited amount of metabolite data that contain in the existed database become a big challenge for researchers to determine biological activities of these volatile metabolites. In previous study, we have accumulated 341 volatiles emitted by biological species associated with 11 types of biological activities and deposited the data into our database, which is called KNApSAcK Metabolite Ecology Database. In this study, we have extended the database by accumulating more data with total of 619 volatiles with 22 different biological activities. 12 models of Deep Neural Network (DNN) were developed to predict the biological activities of volatile metabolites using H2O package in R software. As a comparison, the performance of deep neural network was also being compared with Random Forest (RF) and Gradient Boosting Machine (GBM). Based on the experimental results, DNN6 show the best performance with lower MSE (0.2746) and highest accuracy (76.71%) compared to other models in performance of cross validation. Thus, it indicates that DNN has best accuracy than others. Therefore, DNN method is recommended to be used in the prediction of biological activities of the volatile metabolites. The prediction outcome may be useful for the discovery of novel agricultural tools and biomarkers in medical diagnostic field.

View
Thumbnail Image
Publication

Detection of Parkinson’s Disease (PD) based on speech recordings using machine learning techniques

2020 , Azian Azamimi Abdullah , Nurul Nurain Norazman , Wan Khairunizam Wan Ahmad , Saidatul Ardeenawatie Awang , Foong Wei Jian

There are some neurodegenerative diseases which are unable to cure such as Parkinson's disease (PD) and Hungtinton's disease due to the death of certain parts in the brain that is affecting older adult. PD is an appalling neurodegenerative health disorder that linked to the nervous system which exert influence on motor functions. PD also often known as idiopathic disorder, environmental and genetic factors related, and the causes of PD remain unidentified. To diagnose PD, the clinicians are required to take the history of brain condition for the patient and undergoes various of motor skills examination. Accurate detection of PD plays a crucial role in aiding and providing proper treatment to the patients. Nowadays, there has been recent interest in studying speech-based PD diagnosis. Extracted acoustic attributes are the most important requirement to predict the PD. The experiment was conducted on speech recording dataset consisting of 240 samples. This work studies on the feature selection method, Least Absolute Shrinkage and Selection Operator (LASSO) with multiple machine learnings such as Random Forest (RF), Deep Neural Network (DNN), Gradient Boosting Machine (GBM) and Support Vector Machine (SVM) as the classifier. Throughout this research, train test split method and k-fold cross validation were implemented to evaluate the performance of the classifiers. Through LASSO, Support Vector Machine Grid Search Cross Validation (SVM GSCV) outperformed other 7 models with 100.00 % accuracy, 97.87 % for recall, 65.00 % for specificity and 97.10 % of AUC for 10-fold cross validation. Finally, Graphical User Interface (GUI) was developed and validated through the prediction over UCI speech recording dataset which achieved 96.67 % accuracy for binary classification with 30 samples.

View
Thumbnail Image
Publication

Diagnosis of Heart Disease Using Machine Learning Methods

2021-01-01 , Azian Azamimi Abdullah , Alhadi N.A. , Wan Khairunizam Wan Ahmad

The World Health Organization (WHO) estimated 12 million deaths around the world appear each year from heart disease. Heart disease includes coronary artery disease, heart rhythm problem and heart defects. Each disease has similar symptoms but cause different effects and severity on patient. The common factors of heart disease include high blood pressure, diabetes, cholesterol and age. These factors are independent of each other; thus, the use of artificial intelligence and machine learning will be a suitable choice to model them. Correct diagnosis of heart disease is difficult due to the complicated processes and different system and it is vital because heart disease can lead to a heart attack, chest pain, stroke and sudden death. Hence, an accurate and early detection of heart disease with proper and adequate treatment is needed. The main aim of this research is to identify suitable feature selection method and machine learning algorithms for the diagnosis of heart disease. Chi Square Feature Selection (CSFS), Random Forest Feature Selection (RFFS), Forward Feature Selection (FFS), Backward Feature Selection (BFS) and Exhaustive Feature Selection (EFS) are the feature selection methods applied in this research. These feature selection methods are then implemented in the machine learning algorithms including Random Forest (RF), Support Vector Machine (SVM), Logistic Regression (LR) and k-Nearest Neighbour (KNN). The performance of these machine learning algorithms is evaluated through accuracy, sensitivity and specificity based on Confusion Matrix, ROC Curve and area under ROC (AUC). Based on the results, combination of RF with RFFS produced the highest accuracy value with 85.25% accuracy.

View
Thumbnail Image
Publication

Investigation of a real-time driver eye-closeness for the application of drowsiness detection

2021-01-01 , Kamazlan M.Z.B. , Khairunizam W. , Halin A.H. , Nor M.R.M. , Azian Azamimi Abdullah , Mokhtar N.

The increase in accident and death rates due to drowsiness while driving raises concerns to the community. An efficient solution is vital to ensure the safety of all drivers on the road. Most previous studies have analyzed drowsiness using head tilt, yawning, and eye condition. Face detection applied in drowsiness detection from previous research not included distances between subject and camera. The features used for eye detection required large storage and long-term process which are not applicable in a real-time system. This study uses Haar algorithm and analysis is performed based on the size of the region of interest for face detection. Eye monitoring uses facial landmark features and the evaluation is dependent on the width of the eye. The percentage of eye closure is used to describe the eyes as closed. This study only takes into account the normal rate of blinking eyes while driving because of the long-time constraints required for a person to be in a drowsy state. In this research, the Raspberry Pi 3B+ and Pi cameras are used as processing and vision devices. The highest accuracy of face detection achieved based on the ROI area at a distance of 80 cm is 98.33%. The lowest difference between eye width and the intercanthal distance is 0.36%. The overall normal eye blink rate while driving is in the range of the normal eye blink rate which does not exceed 20 blinks/min as reported by the previous researcher.

View
Thumbnail Image
Publication

Development of Computer Aided System for Classification of Gastrointestinal Lesions

2022-01-01 , Azian Azamimi Abdullah , Kamardin N.A.A. , Wan Khairunizam Wan Ahmad

Colorectal cancer is a major global health problem and is one of the major contributors to deaths worldwide. Because malignant transformations are rare, endoscopic resection of hyperplastic polyps exacerbates medical costs, including those for resection and unnecessary pathological assessment. The proposed work is based on assessment of exploratory data analysis and visualization, initial pre-processing step followed by selection of attributes and performance assessment of proposed supervised machine learning algorithms. For features selection, two method were implemented (Boruta and SelectFromModel (SFM) Random Forest) to compare the performances of the models. The comparative analysis of machine learning included Random Forest (RF), Support Vector Machines (SVM), Deep Neural Network (DNN). For Boruta algorithm, it is shown that RF has the highest accuracy of 90.32%, sensitivity of 93.05% and specificity of 95.95%. Therefore, the adenoma detection rate (ADR) is desirable for improvement. A computer-aided system is developed which offers the opportunity to evaluate the presence of colorectal polyps objectively during colonoscopy.

View
Thumbnail Image
Publication

Detection of Parkinson's Disease (PD) Based on Speech Recordings using Machine Learning Techniques

2020-12-20 , Azian Azamimi Abdullah , Norazman N.N. , Wan Khairunizam Wan Ahmad , Saidatul Ardeenawatie Awang , Jian F.W.

There are some neurodegenerative diseases which are unable to cure such as Parkinson's disease (PD) and Hungtinton's disease due to the death of certain parts in the brain that is affecting older adult. PD is an appalling neurodegenerative health disorder that linked to the nervous system which exert influence on motor functions. PD also often known as idiopathic disorder, environmental and genetic factors related, and the causes of PD remain unidentified. To diagnose PD, the clinicians are required to take the history of brain condition for the patient and undergoes various of motor skills examination. Accurate detection of PD plays a crucial role in aiding and providing proper treatment to the patients. Nowadays, there has been recent interest in studying speech-based PD diagnosis. Extracted acoustic attributes are the most important requirement to predict the PD. The experiment was conducted on speech recording dataset consisting of 240 samples. This work studies on the feature selection method, Least Absolute Shrinkage and Selection Operator (LASSO) with multiple machine learnings such as Random Forest (RF), Deep Neural Network (DNN), Gradient Boosting Machine (GBM) and Support Vector Machine (SVM) as the classifier. Throughout this research, train test split method and k-fold cross validation were implemented to evaluate the performance of the classifiers. Through LASSO, Support Vector Machine Grid Search Cross Validation (SVM GSCV) outperformed other 7 models with 100.00 % accuracy, 97.87 % for recall, 65.00 % for specificity and 97.10 % of AUC for 10-fold cross validation. Finally, Graphical User Interface (GUI) was developed and validated through the prediction over UCI speech recording dataset which achieved 96.67 % accuracy for binary classification with 30 samples.

View
Thumbnail Image
Publication

Effect of Tryptone Concentration on Cyclodextrin Glucanotranferase (CGTase) excretion and cell lysis of immobilized recombinant Escherichia coli

2020 , R C Man , R M Illias , S M Shaarani , Z I M Arshad , S K A Mudalip , S Z Sulaiman , S F Z Mohamad Fuzi , Azian Azamimi Abdullah

The recombinant enzyme excretion into the culture medium provides significant advantages over cytoplasmic expression. Nevertheless, the problems encountered during the excretion of recombinant enzyme are the plasmid instability and occurrence of cell lysis. Various attempts have been made to improve the recombinant enzyme excretion and plasmid stability with the low occurrence of cell lysis. The approaches include the modification of the nitrogen sources in the medium such as tryptone, the use of cell immobilization system and lowering the induction temperature. In the present study, the effects of different tryptone concentrations (1, 5, 10, 20 and 30 g/L) as nitrogen source in super optimal broth (SOB) medium on CGTase excretion and plasmid stability as well as cell lysis of the immobilized cell were studied. The recombinant E. coli was immobilized on polyvinylidene fluoride polymer (PVDF) hollow fiber membrane. The immobilized cells were expressed using 0.011 mM IPTG at 25°C, 200 rpm of agitation rate and pH 8.8 for 24 h of post induction time. The use of low tryptone concentration (5 g/l) produced high CGTase excretion (758.64 U/ml) and increased the plasmid stability (86% increment) with reduction of cell lysis (90% reduction) in comparison with the initial tryptone concentration (20 g/l). Hence, low concentration of tryptone could reduce the cost for CGTase production due to low amount of tryptone used in the fermentation process.

View
Thumbnail Image
Publication

Analysis of familiar and unfamiliar images using power spectral estimation for EEG authentication system

2021-01-01 , Rosli F.A. , Saidatul A. , Azian Azamimi Abdullah , Hilmi A.H.

Biometric authentication is a system used for recognizing an individual according to their physiological and behavioral characteristics, Recently, the application of biometric authentication is the most useful in cybersecurity such as fingerprints, facial, and voices. Traditional authentication such as password and PIN have been used for a decade, however, they bring drawbacks to the users which were attacked by cybercriminals. Therefore, the brainwave of electroencephalogram (EEG) is proposed as the biometric trait to encounter the problems faced. The aim of this study is to explore the feature extraction method by applying the power spectral estimation method as linear feature analysis such as the Welch method, Burg Method, Yule-Walker method, Covariance method and Modified Covariance method. After extracting five features, the statistics of mean, median, variance, and the standard deviation is computed and fed into three types of shallow classifiers including the Neural Network, K-Nearest Neighbors, and Support Vector Machine. As a result, the Yule-Walker feature contributes to the highest average accuracy and the Neural Network using Levenberg-Marquardt (LM) has achieved the highest accuracy for all frequency bands. In fact, the highest frequency band obtained by gamma (30-45 Hz) followed by highbeta (20-30 Hz), lowbeta (13-20 Hz) and alpha (8-13 Hz). Overall, all three features and three classifiers are able to achieve between 70.1% to 98.2% of accuracy which shows that it can differentiate between different tasks.

View
Thumbnail Image
Publication

Recent Trends in Biomedical Engineering and Healthcare

2019 , Azian Azamimi Abdullah , Mohd Hanafi Mat Som , Zulkarnay Zakaria

The field of biomedical engineering has emerged rapidly with new technologies that aim to improve the quality of life and healthcare. This field seeks to close the gap between engineering and medicine, combining the design and problem-solving skills of engineering with medical biological sciences to advance healthcare treatment, including diagnosis, monitoring, and therapy. This book focuses on the state-of-the-art and recent trends in the emerging field of biomedical engineering. The topics such as biomedical signal and image processing, bio instrumentations, biomedical electronics and devices, biomaterials, biomechanics medical imaging, augmented reality, and bioinformatics are widely covered in this book. The present book is developed to make the reader familiar with the technologies that can improve human health. It will cater to the graduate and post-graduate students of electronic, mechanical, mechatronic, and biomedical engineering. This book is also suitable for researchers and practitioners who have an interest in recent trends in biomedical engineering and healthcare.

View