Thulasyammal Ramiah Pillai - Profile on Academia.edu (original) (raw)

Papers by Thulasyammal Ramiah Pillai

Air pollution is one of the significant causes of mortality and morbidity every year. In recent y... more Air pollution is one of the significant causes of mortality and morbidity every year. In recent years, many researchers have focused their attention on the associations of air pollution and health. Air pollution data and health data is used in these studies and feature engineering is used to create and optimize the air quality and health features. In order to associate these datasets, the residential address, community/county/block/city, and hospital/school address are utilized as association parameters. A spatial problem is raised when the Air Quality Monitoring (AQM) stations are concentrated in urban areas within the regions, and the residential address or any other spatial parameter is used. An intersection of AQM stations coverage in urban areas is observed where AQM stations are operating in close proximity, which raises the question of how to associate the patients with the relevant AQM station. In most studies, the distance of patients to the AQM stations is also not taken into account. In this study, we propose a spatial feature engineering algorithm with functions to find the coordinates for patients, calculate distances to the AQM stations, and associate patient records to the nearest AQM station. Hence, removing the limitations of current air pollution health datasets. The proposed algorithm is applied to a case study in Klang Valley, Malaysia. The results show that the proposed algorithm can generate air pollution health datasets efficiently, and it also provides the radius facility to exclude the patients who are situated far away from the stations.

Application of GARMA(1,1;1,d) model to GDP in Malaysia: An illustrative example

Gross Domestic Product (GDP) per capita is often used as an indicator of standard of living in an... more Gross Domestic Product (GDP) per capita is often used as an indicator of standard of living in an economy. GDP per capita observed over the years can be modelled using time series models. A new class of GARMA has been introduced in the time series literature to reveal some hidden features in time series data. In this paper, we illustrate the fitting of GARMA (1, 1; 1,) model to the GDP growth data of Malaysia which has been observed from 1955 to 2009. The estimation of the model was done using Hannan-Rissanen Algorithm

Artificial intelligence techniques for predicting cardiorespiratory mortality caused by air pollution

International Journal of Environmental Science and Technology, Apr 5, 2022

Air pollution (AP) has risen as one of the biggest challenges of the 21st century, and it has adv... more Air pollution (AP) has risen as one of the biggest challenges of the 21st century, and it has adverse health effects for humans. The effects of health effects, including cardiorespiratory health effects of various air pollutants, are well documented. This research work presents the modeling and analysis of cardiorespiratory mortality attributed to AP. The modeling and predictions are also completed using four Artificial Intelligence (AI) techniques for comparison. The AI techniques utilized for comparison are (1) Enhanced Long Short-Term Memory (ELSTM), (2) Vector Autoregressive (VAR) (3) Deep Learning (DL), and (4) Long Short-Term Memory (LSTM). The research work is carried out at seven locations in Klang Valley, Malaysia. The five study locations i.e., Cheras, Petaling Jaya, Putrajaya, Shah Alam, and Klang have data from January 2006 to December 2016 and two relatively new monitoring stations, i.e., Banting and Batu Muda have data from April 2010 to December 2016 and January 2009 to December 2016, respectively. The comparison of results indicates that the ELSTM model predicts the cardiorespiratory mortality caused by AP significantly better than other AI models utilized in the study. The best Root-Mean-Squared Error (RMSE) results are obtained at Batu Muda and Klang study locations (ELSTM: 0.004, VAR: 0.03, DL: 0.0081, LSTM: 0.006) and (ELSTM: 0.005, VAR: 0.114, DL: 0.076, LSTM: 0.020), respectively. Based on the results, we can conclude that we can predict cardiorespiratory mortality based on air pollution in Klang Valley, Malaysia, using the AI techniques utilized in the study, especially ELSTM.

Credit Card Fraud Detection Using Deep Learning Technique

Credit card fraud detection is growing due to the increase and the popularity of online banking. ... more Credit card fraud detection is growing due to the increase and the popularity of online banking. The need to detect fraudulent within credit card has become as a serious problem among the online shoppers. The multi-layer perceptron (MLP) machine learning algorithm is used to identify the credit card fraud. We have used the various parameters of the MLP to compare the performance of MLP. The aim of this paper is to design a high performance model to detect the credit card fraud using deep learning techniques. We found that logistic and hyperbolic tangent activation function offer good performance in detecting the credit card fraud. The logistic activation function performs better when there are 10 nodes, the sensitivity is 82% and when there are 100 nodes, the sensitivity is 83% respectively in the 3 hidden layer model. However, hyperbolic tangent activation function performs better when there is 1000 nodes, the sensitivity is 82% in all the number (1, 2 and 3) of hidden layers. This study will give us a guidance on how to choose a best model to obtain optimum results with minimum cost in deep learning.

Sensors, 2021

Exposure to mental stress for long period leads to serious accidents and health problems. To avoi... more Exposure to mental stress for long period leads to serious accidents and health problems. To avoid negative consequences on health and safety, it is very important to detect mental stress at its early stages, i.e., when it is still limited to acute or episodic stress. In this study, we developed an experimental protocol to induce two different levels of stress by utilizing a mental arithmetic task with time pressure and negative feedback as the stressors. We assessed the levels of stress on 22 healthy subjects using frontal electroencephalogram (EEG) signals, salivary alpha-amylase level (AAL), and multiple machine learning (ML) classifiers. The EEG signals were analyzed using a fusion of functional connectivity networks estimated by the Phase Locking Value (PLV) and temporal and spectral domain features. A total of 210 different features were extracted from all domains. Only the optimum multi-domain features were used for classification. We then quantified stress levels using stati...

Research Square (Research Square), Apr 5, 2021

Air pollution has a serious and adverse effect on human health, and it has become a risk to human... more Air pollution has a serious and adverse effect on human health, and it has become a risk to human welfare and health throughout the globe. In this paper, we present the modeling and analysis of air pollution and cardiorespiratory hospitalization. This study aims to investigate the association between cardiorespiratory hospitalization and air pollution, and predict cardiorespiratory hospitalization based on air pollution using the Artificial Intelligence (AI) techniques. We propose the Enhanced Long Short-Term Memory (EL-STM) model and provide a comparison with other AI techniques, i.e., Long Short-Term Memory (LSTM), Deep Learning (DL), and Vector Autoregressive (VAR). This study was conducted at seven study locations in Klang Valley, Malaysia. The prediction results show that the ELSTM model performed significantly better than other models in all study locations, with the best RMSE scores in Klang study location (ELSTM: 0.002, LSTM: 0.013, DL: 0.006, VAR: 0.066). The results also indicated that the proposed ELSTM model was able

A Multilayer Perceptron Model for the Classification of Breast Cancer Cells

Application of land use regression model to assess outdoor air pollution exposure: A review

Environmental Advances

A wearable single EEG channel analysis for mental stress state detection

2021 IEEE 7th International Conference on Computing, Engineering and Design (ICCED), 2021

Mental stress is a world’s apprising issue due to its impact on health and the economy. Chronic s... more Mental stress is a world’s apprising issue due to its impact on health and the economy. Chronic stress negatively affects human cognitive abilities and decision-making. To avoid its serious consequences, it is paramount important to detect it at an early stage. In this study, we assessed the levels of stress on 28 healthy subjects by utilizing an Electroencephalogram (EEG) of a single channel and machine learning approach. The EEG signals were analyzed by extracting 20 features from the time and frequency domains. The optimum features were, then, selected using decision trees of information gain. Consequently, we classified the levels of stress using support vector machines (SVM) classifier with a GRID Search optimizer. The proposed feature selection method results in a 66% reduction of feature vector space and achieved an accuracy of 86% using the optimized SVM classifier. Our result demonstrates the effectiveness of the proposed method for the development of real-life stress applications.

Mental stress state recognition using electroencephalogram (EEG) signals for real-life applicatio... more Mental stress state recognition using electroencephalogram (EEG) signals for real-life applications needs a conventional wearable device. This requires an efficient number of EEG channels and an optimal feature set. The main objective of the study is to identify an optimal feature subset that can best discriminate mental stress states while enhancing the overall performance. Thus, multi-domain feature extraction methods were employed, namely, time domain, frequency domain, time-frequency domain, and network connectivity features, to form a large feature vector space. To avoid the computational complexity of high dimensional space, a hybrid feature selection (FS) method of minimum Redundancy Maximum Relevance with Particle Swarm Optimization and Support Vector Machine (mRMR-PSO-SVM) is proposed to remove noise, redundant, and irrelevant features and keep the optimal feature subset. The performance of the proposed method is evaluated and verified using four datasets, namely EDMSS, DEA...

Smart Mobility Cities: Connecting Bristol and Kuala Lumpur

Financed by the British Council Institutional Links program this Smart Mobility Cities project ha... more Financed by the British Council Institutional Links program this Smart Mobility Cities project has opened a fascinating window on a journey of discovery linking Bristol and Kuala Lumpur. This journey was in part directed towards the realisation of Smart Mobility solutions to the socio-economic and environmental challenges of global urbanisation. Beyond this, the journey was also concerned to strengthen research and innovation partnerships between the UK and the emerging knowledge economy of Malaysia, enabling UK social scientists to collaborate on challenging global issues with international researchers and vice versa. This Smart Mobility Cities project report presents innovative, creative and yet fully practical solutions for these societal challenges. Solutions that explore a range of opportunities, which include those arising from new urban governance requirements, and which are in-line with visions for sustainable urban mobility. These Smart Mobility solutions have arisen from intensive co-design and co-creation engagement with a diversity of stakeholders. Research co-production has linked the principal university partners of the University of the West of England (UWE), Bristol, and Taylor’s University, Kuala Lumpur, together with the Malaysia Institute of Transport (MITRANS), Universiti Teknologi Mara, and the University Sains Malaysia (USM) in intensive engagement with stakeholder interests in both UK and Malaysia over a two-year period.

Application of Garma (1, 1; 1, &) Model to GDP in Malaysia: An Illustrative Example

Gross Domestic Product (GDP) per capita is often used as an indicator of standard of living in an... more Gross Domestic Product (GDP) per capita is often used as an indicator of standard of living in an economy. GDP per capita observed over the years can be modelled using time series models. A new class of GARMA has been introduced in the time series literature to reveal some hidden features in time series data. In this paper, we illustrate the fitting of GARMA (1, 1; 1,) model to the GDP growth data of Malaysia which has been observed from 1955 to 2009. The estimation of the model was done using Hannan-Rissanen Algorithm.

Gross Domestic Product (GDP) per capita is often used as an indicator of standard of living in an... more Gross Domestic Product (GDP) per capita is often used as an indicator of standard of living in an economy. GDP per capita observed over the years can be modelled using time series models. A new class of Generalized Autoregressive Moving Average (GARMA) namely GARMA (1, 2; δ, 1) has been introduced in the time series literature to reveal some hidden features in time series. In this paper, GARMA (1, 2;δ, 1) model and ARMA (1, 1) model are fitted in the GDP growth data of Malaysia which has been observed from 1955 to 2009. The parameter estimation methods considered include the Hannan Rissanen Algorithm (HRA), Whittle Estimation (WE) and Maximum Likelihood Estimation (MLE). Point forecasts also have been done and the performance of GARMA (1, 2; δ, 1) and ARMA (1, 1) and the estimation methods are discussed.

Time Series Properties of the Class of First Order Autoregressive Processes with Generalized Moving Average Errors

A new class of time series models known as Generalized Autoregressive of order one with first ord... more A new class of time series models known as Generalized Autoregressive of order one with first order moving average errors has been introduced in order to reveal some features of certain time series data. The variance and autocovariance of the

GARMA Modeling of ECG and Classification of Arrhythmia

2018 8th International Conference on Intelligent Systems, Modelling and Simulation (ISMS), 2018

Computer-assisted arrhythmia detection is crucial for the treatment of cardiac disorders. Electro... more Computer-assisted arrhythmia detection is crucial for the treatment of cardiac disorders. Electrocardiograms (ECG) are used to study the electric heart activity and diagnose abnormalities in the heart. It is a non-invasive method where the electric signal of the heart is captured through electrodes placed on the skin. In this study the ECG signals will be classified according to the heart abnormalities using the Generalized Autoregressive Moving Average (GARMA) and Generalized Linear Model (GLM). The ECG features were extracted using the GARMA model. The coefficients obtained from GARMA model will be classified using the GLM model to detect and classify into normal and five types of arrhythmias.

Comparative Analysis of ARMA and GARMA Models in Forecasting

In this paper, two traditional Autoregressive Moving Average models and two different Generalised... more In this paper, two traditional Autoregressive Moving Average models and two different Generalised Autoregressive Moving Average models are considered to forecast financial time series. These time series models are fitted to the financial time series data namely Dow Jones Utilities Index data set, Daily Closing Value of the Dow Jones Average and Daily Returns of the Dow Jones Utilities Average Index. Three different estimation methods such as Hannan-Rissanen Algorithm, Whittle’s Estimation and Maximum Likelihood Estimation are used to estimate the parameters of the models. Point forecasts have been done and the performance of all the models and the estimation methods are discussed.

Accuracy Comparison of Machine Learning Algorithms for Predictive Analytics in Higher Education

Lecture Notes of the Institute for Computer Sciences, Social Informatics and Telecommunications Engineering, 2019

In this research, we compared the accuracy of machine learning algorithms that could be used for ... more In this research, we compared the accuracy of machine learning algorithms that could be used for predictive analytics in higher education. The proposed experiment is based on a combination of classic machine learning algorithms such as Naive Bayes and Random Forest with various ensemble methods such as Stochastic, Linear Discriminant Analysis (LDA), Tree model (C5.0), Bagged CART (treebag) and K Nearest Neighbors (KNN). We applied traditional classification methods to classify the students’ performance and to determine the independent variables that offer the highest accuracy. Our results depict that the data with the 11 features using random forest generated the best accuracy value of 0.7333. However, we revised the experiment with ensemble algorithms to reduce the variance (bagging), bias (boosting) and to improve the prediction accuracy (stacking). Consequently, the bagging random forest outperformed other methods with the accuracy value of 0.7959.

Lecture Notes of the Institute for Computer Sciences, Social Informatics and Telecommunications Engineering, 2019

The journey of data from the state of being valueless to valuable has been possible due to powerf... more The journey of data from the state of being valueless to valuable has been possible due to powerful analytics tools and processing platforms. Organizations have realized the potential of data, and they are looking far ahead from the traditional relational databases to unstructured as well as semi-structured data generated from heterogeneous sources. With the numerous devices and sensors surrounding our ecosystem, IoT has become a reality, and with the use of data science, IoT analytics has become a tremendous opportunity to perceive incredible insights. However, despite the various benefits of IoT analytics, organizations are apprehensive with the dark side of IoT such as security and privacy concerns. In this research, we discuss the opportunities and concerns of IoT analytics. Moreover, we propose a generic data science methodology for IoT data analytics named as Plan, Collect and Analytics for Internet-of-Things (PCA-IoT). The proposed methodology could be applied in IoT scenario...

Properties of selected garma models and their estimation procedures

Time series is an ordered sequence of random variables. In other words, a time series is a set of... more Time series is an ordered sequence of random variables. In other words, a time series is a set of observations fxtg, each one being recorded at a speci¯c time t. Usually time series are modelled as Autoregressive Moving Average (ARMA), Autoregressive Integrated Moving Average (ARIMA), Autoregressive Fractional Integrated Moving Average (ARFIMA) and etc. An extension of the class of time series models by introducing a new parameter, ± as an index includes Generalized Autoregressive (GAR(p)), Generalized Moving Average (GMA(q)) and Generalized Autoregressive Moving Average (GARMA (p; q; ±1; ±2)). The focus of this study is to investigate the properties specically the variance and autocovariance of the GARMA (p; q; ±1; ±2) models. We also study the estimation of the parameters of these models. Evaluation of the performance of two estimators based on the Hannan-Rissanen Algorithm Estimator (HRA) and the Whittle's Estimator (WE) through a series of simulation studies have been conduc...

A comparison of Estimation Techniques for electrocardiogram classification

Electrocardiograms are one of the most important bio-signals used to diagnose heart conditions su... more Electrocardiograms are one of the most important bio-signals used to diagnose heart conditions such as arrhythmia. In this paper we interpret the rhythm and identify the potential disruptions using computer assisted methods of interpretation. Moreover, Hannan-Rissanen algorithm (HRA) and Maximum Likelihood (MLE) are used to estimate the coefficients of the Generalized Autoregressive Moving Average (GARMA). We classified ECG signals into 6 classes (normal and 5 types of arrhythmia) using neural networks approach. We also compared the results of HRA and MLE estimations methods and found that HRA performs better than MLE. The overall detection accuracy using HRA is 81.7% and 67.3% for MLE.

Air pollution is one of the significant causes of mortality and morbidity every year. In recent y... more Air pollution is one of the significant causes of mortality and morbidity every year. In recent years, many researchers have focused their attention on the associations of air pollution and health. Air pollution data and health data is used in these studies and feature engineering is used to create and optimize the air quality and health features. In order to associate these datasets, the residential address, community/county/block/city, and hospital/school address are utilized as association parameters. A spatial problem is raised when the Air Quality Monitoring (AQM) stations are concentrated in urban areas within the regions, and the residential address or any other spatial parameter is used. An intersection of AQM stations coverage in urban areas is observed where AQM stations are operating in close proximity, which raises the question of how to associate the patients with the relevant AQM station. In most studies, the distance of patients to the AQM stations is also not taken into account. In this study, we propose a spatial feature engineering algorithm with functions to find the coordinates for patients, calculate distances to the AQM stations, and associate patient records to the nearest AQM station. Hence, removing the limitations of current air pollution health datasets. The proposed algorithm is applied to a case study in Klang Valley, Malaysia. The results show that the proposed algorithm can generate air pollution health datasets efficiently, and it also provides the radius facility to exclude the patients who are situated far away from the stations.

Application of GARMA(1,1;1,d) model to GDP in Malaysia: An illustrative example

Gross Domestic Product (GDP) per capita is often used as an indicator of standard of living in an... more Gross Domestic Product (GDP) per capita is often used as an indicator of standard of living in an economy. GDP per capita observed over the years can be modelled using time series models. A new class of GARMA has been introduced in the time series literature to reveal some hidden features in time series data. In this paper, we illustrate the fitting of GARMA (1, 1; 1,) model to the GDP growth data of Malaysia which has been observed from 1955 to 2009. The estimation of the model was done using Hannan-Rissanen Algorithm

Artificial intelligence techniques for predicting cardiorespiratory mortality caused by air pollution

International Journal of Environmental Science and Technology, Apr 5, 2022

Air pollution (AP) has risen as one of the biggest challenges of the 21st century, and it has adv... more Air pollution (AP) has risen as one of the biggest challenges of the 21st century, and it has adverse health effects for humans. The effects of health effects, including cardiorespiratory health effects of various air pollutants, are well documented. This research work presents the modeling and analysis of cardiorespiratory mortality attributed to AP. The modeling and predictions are also completed using four Artificial Intelligence (AI) techniques for comparison. The AI techniques utilized for comparison are (1) Enhanced Long Short-Term Memory (ELSTM), (2) Vector Autoregressive (VAR) (3) Deep Learning (DL), and (4) Long Short-Term Memory (LSTM). The research work is carried out at seven locations in Klang Valley, Malaysia. The five study locations i.e., Cheras, Petaling Jaya, Putrajaya, Shah Alam, and Klang have data from January 2006 to December 2016 and two relatively new monitoring stations, i.e., Banting and Batu Muda have data from April 2010 to December 2016 and January 2009 to December 2016, respectively. The comparison of results indicates that the ELSTM model predicts the cardiorespiratory mortality caused by AP significantly better than other AI models utilized in the study. The best Root-Mean-Squared Error (RMSE) results are obtained at Batu Muda and Klang study locations (ELSTM: 0.004, VAR: 0.03, DL: 0.0081, LSTM: 0.006) and (ELSTM: 0.005, VAR: 0.114, DL: 0.076, LSTM: 0.020), respectively. Based on the results, we can conclude that we can predict cardiorespiratory mortality based on air pollution in Klang Valley, Malaysia, using the AI techniques utilized in the study, especially ELSTM.

Credit Card Fraud Detection Using Deep Learning Technique

Credit card fraud detection is growing due to the increase and the popularity of online banking. ... more Credit card fraud detection is growing due to the increase and the popularity of online banking. The need to detect fraudulent within credit card has become as a serious problem among the online shoppers. The multi-layer perceptron (MLP) machine learning algorithm is used to identify the credit card fraud. We have used the various parameters of the MLP to compare the performance of MLP. The aim of this paper is to design a high performance model to detect the credit card fraud using deep learning techniques. We found that logistic and hyperbolic tangent activation function offer good performance in detecting the credit card fraud. The logistic activation function performs better when there are 10 nodes, the sensitivity is 82% and when there are 100 nodes, the sensitivity is 83% respectively in the 3 hidden layer model. However, hyperbolic tangent activation function performs better when there is 1000 nodes, the sensitivity is 82% in all the number (1, 2 and 3) of hidden layers. This study will give us a guidance on how to choose a best model to obtain optimum results with minimum cost in deep learning.

Sensors, 2021

Exposure to mental stress for long period leads to serious accidents and health problems. To avoi... more Exposure to mental stress for long period leads to serious accidents and health problems. To avoid negative consequences on health and safety, it is very important to detect mental stress at its early stages, i.e., when it is still limited to acute or episodic stress. In this study, we developed an experimental protocol to induce two different levels of stress by utilizing a mental arithmetic task with time pressure and negative feedback as the stressors. We assessed the levels of stress on 22 healthy subjects using frontal electroencephalogram (EEG) signals, salivary alpha-amylase level (AAL), and multiple machine learning (ML) classifiers. The EEG signals were analyzed using a fusion of functional connectivity networks estimated by the Phase Locking Value (PLV) and temporal and spectral domain features. A total of 210 different features were extracted from all domains. Only the optimum multi-domain features were used for classification. We then quantified stress levels using stati...

Research Square (Research Square), Apr 5, 2021

Air pollution has a serious and adverse effect on human health, and it has become a risk to human... more Air pollution has a serious and adverse effect on human health, and it has become a risk to human welfare and health throughout the globe. In this paper, we present the modeling and analysis of air pollution and cardiorespiratory hospitalization. This study aims to investigate the association between cardiorespiratory hospitalization and air pollution, and predict cardiorespiratory hospitalization based on air pollution using the Artificial Intelligence (AI) techniques. We propose the Enhanced Long Short-Term Memory (EL-STM) model and provide a comparison with other AI techniques, i.e., Long Short-Term Memory (LSTM), Deep Learning (DL), and Vector Autoregressive (VAR). This study was conducted at seven study locations in Klang Valley, Malaysia. The prediction results show that the ELSTM model performed significantly better than other models in all study locations, with the best RMSE scores in Klang study location (ELSTM: 0.002, LSTM: 0.013, DL: 0.006, VAR: 0.066). The results also indicated that the proposed ELSTM model was able

A Multilayer Perceptron Model for the Classification of Breast Cancer Cells

Application of land use regression model to assess outdoor air pollution exposure: A review

Environmental Advances

A wearable single EEG channel analysis for mental stress state detection

2021 IEEE 7th International Conference on Computing, Engineering and Design (ICCED), 2021

Mental stress is a world’s apprising issue due to its impact on health and the economy. Chronic s... more Mental stress is a world’s apprising issue due to its impact on health and the economy. Chronic stress negatively affects human cognitive abilities and decision-making. To avoid its serious consequences, it is paramount important to detect it at an early stage. In this study, we assessed the levels of stress on 28 healthy subjects by utilizing an Electroencephalogram (EEG) of a single channel and machine learning approach. The EEG signals were analyzed by extracting 20 features from the time and frequency domains. The optimum features were, then, selected using decision trees of information gain. Consequently, we classified the levels of stress using support vector machines (SVM) classifier with a GRID Search optimizer. The proposed feature selection method results in a 66% reduction of feature vector space and achieved an accuracy of 86% using the optimized SVM classifier. Our result demonstrates the effectiveness of the proposed method for the development of real-life stress applications.

Mental stress state recognition using electroencephalogram (EEG) signals for real-life applicatio... more Mental stress state recognition using electroencephalogram (EEG) signals for real-life applications needs a conventional wearable device. This requires an efficient number of EEG channels and an optimal feature set. The main objective of the study is to identify an optimal feature subset that can best discriminate mental stress states while enhancing the overall performance. Thus, multi-domain feature extraction methods were employed, namely, time domain, frequency domain, time-frequency domain, and network connectivity features, to form a large feature vector space. To avoid the computational complexity of high dimensional space, a hybrid feature selection (FS) method of minimum Redundancy Maximum Relevance with Particle Swarm Optimization and Support Vector Machine (mRMR-PSO-SVM) is proposed to remove noise, redundant, and irrelevant features and keep the optimal feature subset. The performance of the proposed method is evaluated and verified using four datasets, namely EDMSS, DEA...

Smart Mobility Cities: Connecting Bristol and Kuala Lumpur

Financed by the British Council Institutional Links program this Smart Mobility Cities project ha... more Financed by the British Council Institutional Links program this Smart Mobility Cities project has opened a fascinating window on a journey of discovery linking Bristol and Kuala Lumpur. This journey was in part directed towards the realisation of Smart Mobility solutions to the socio-economic and environmental challenges of global urbanisation. Beyond this, the journey was also concerned to strengthen research and innovation partnerships between the UK and the emerging knowledge economy of Malaysia, enabling UK social scientists to collaborate on challenging global issues with international researchers and vice versa. This Smart Mobility Cities project report presents innovative, creative and yet fully practical solutions for these societal challenges. Solutions that explore a range of opportunities, which include those arising from new urban governance requirements, and which are in-line with visions for sustainable urban mobility. These Smart Mobility solutions have arisen from intensive co-design and co-creation engagement with a diversity of stakeholders. Research co-production has linked the principal university partners of the University of the West of England (UWE), Bristol, and Taylor’s University, Kuala Lumpur, together with the Malaysia Institute of Transport (MITRANS), Universiti Teknologi Mara, and the University Sains Malaysia (USM) in intensive engagement with stakeholder interests in both UK and Malaysia over a two-year period.

Application of Garma (1, 1; 1, &) Model to GDP in Malaysia: An Illustrative Example

Gross Domestic Product (GDP) per capita is often used as an indicator of standard of living in an... more Gross Domestic Product (GDP) per capita is often used as an indicator of standard of living in an economy. GDP per capita observed over the years can be modelled using time series models. A new class of GARMA has been introduced in the time series literature to reveal some hidden features in time series data. In this paper, we illustrate the fitting of GARMA (1, 1; 1,) model to the GDP growth data of Malaysia which has been observed from 1955 to 2009. The estimation of the model was done using Hannan-Rissanen Algorithm.

Gross Domestic Product (GDP) per capita is often used as an indicator of standard of living in an... more Gross Domestic Product (GDP) per capita is often used as an indicator of standard of living in an economy. GDP per capita observed over the years can be modelled using time series models. A new class of Generalized Autoregressive Moving Average (GARMA) namely GARMA (1, 2; δ, 1) has been introduced in the time series literature to reveal some hidden features in time series. In this paper, GARMA (1, 2;δ, 1) model and ARMA (1, 1) model are fitted in the GDP growth data of Malaysia which has been observed from 1955 to 2009. The parameter estimation methods considered include the Hannan Rissanen Algorithm (HRA), Whittle Estimation (WE) and Maximum Likelihood Estimation (MLE). Point forecasts also have been done and the performance of GARMA (1, 2; δ, 1) and ARMA (1, 1) and the estimation methods are discussed.

Time Series Properties of the Class of First Order Autoregressive Processes with Generalized Moving Average Errors

A new class of time series models known as Generalized Autoregressive of order one with first ord... more A new class of time series models known as Generalized Autoregressive of order one with first order moving average errors has been introduced in order to reveal some features of certain time series data. The variance and autocovariance of the

GARMA Modeling of ECG and Classification of Arrhythmia

2018 8th International Conference on Intelligent Systems, Modelling and Simulation (ISMS), 2018

Computer-assisted arrhythmia detection is crucial for the treatment of cardiac disorders. Electro... more Computer-assisted arrhythmia detection is crucial for the treatment of cardiac disorders. Electrocardiograms (ECG) are used to study the electric heart activity and diagnose abnormalities in the heart. It is a non-invasive method where the electric signal of the heart is captured through electrodes placed on the skin. In this study the ECG signals will be classified according to the heart abnormalities using the Generalized Autoregressive Moving Average (GARMA) and Generalized Linear Model (GLM). The ECG features were extracted using the GARMA model. The coefficients obtained from GARMA model will be classified using the GLM model to detect and classify into normal and five types of arrhythmias.

Comparative Analysis of ARMA and GARMA Models in Forecasting

In this paper, two traditional Autoregressive Moving Average models and two different Generalised... more In this paper, two traditional Autoregressive Moving Average models and two different Generalised Autoregressive Moving Average models are considered to forecast financial time series. These time series models are fitted to the financial time series data namely Dow Jones Utilities Index data set, Daily Closing Value of the Dow Jones Average and Daily Returns of the Dow Jones Utilities Average Index. Three different estimation methods such as Hannan-Rissanen Algorithm, Whittle’s Estimation and Maximum Likelihood Estimation are used to estimate the parameters of the models. Point forecasts have been done and the performance of all the models and the estimation methods are discussed.

Accuracy Comparison of Machine Learning Algorithms for Predictive Analytics in Higher Education

Lecture Notes of the Institute for Computer Sciences, Social Informatics and Telecommunications Engineering, 2019

In this research, we compared the accuracy of machine learning algorithms that could be used for ... more In this research, we compared the accuracy of machine learning algorithms that could be used for predictive analytics in higher education. The proposed experiment is based on a combination of classic machine learning algorithms such as Naive Bayes and Random Forest with various ensemble methods such as Stochastic, Linear Discriminant Analysis (LDA), Tree model (C5.0), Bagged CART (treebag) and K Nearest Neighbors (KNN). We applied traditional classification methods to classify the students’ performance and to determine the independent variables that offer the highest accuracy. Our results depict that the data with the 11 features using random forest generated the best accuracy value of 0.7333. However, we revised the experiment with ensemble algorithms to reduce the variance (bagging), bias (boosting) and to improve the prediction accuracy (stacking). Consequently, the bagging random forest outperformed other methods with the accuracy value of 0.7959.

Lecture Notes of the Institute for Computer Sciences, Social Informatics and Telecommunications Engineering, 2019

The journey of data from the state of being valueless to valuable has been possible due to powerf... more The journey of data from the state of being valueless to valuable has been possible due to powerful analytics tools and processing platforms. Organizations have realized the potential of data, and they are looking far ahead from the traditional relational databases to unstructured as well as semi-structured data generated from heterogeneous sources. With the numerous devices and sensors surrounding our ecosystem, IoT has become a reality, and with the use of data science, IoT analytics has become a tremendous opportunity to perceive incredible insights. However, despite the various benefits of IoT analytics, organizations are apprehensive with the dark side of IoT such as security and privacy concerns. In this research, we discuss the opportunities and concerns of IoT analytics. Moreover, we propose a generic data science methodology for IoT data analytics named as Plan, Collect and Analytics for Internet-of-Things (PCA-IoT). The proposed methodology could be applied in IoT scenario...

Properties of selected garma models and their estimation procedures

Time series is an ordered sequence of random variables. In other words, a time series is a set of... more Time series is an ordered sequence of random variables. In other words, a time series is a set of observations fxtg, each one being recorded at a speci¯c time t. Usually time series are modelled as Autoregressive Moving Average (ARMA), Autoregressive Integrated Moving Average (ARIMA), Autoregressive Fractional Integrated Moving Average (ARFIMA) and etc. An extension of the class of time series models by introducing a new parameter, ± as an index includes Generalized Autoregressive (GAR(p)), Generalized Moving Average (GMA(q)) and Generalized Autoregressive Moving Average (GARMA (p; q; ±1; ±2)). The focus of this study is to investigate the properties specically the variance and autocovariance of the GARMA (p; q; ±1; ±2) models. We also study the estimation of the parameters of these models. Evaluation of the performance of two estimators based on the Hannan-Rissanen Algorithm Estimator (HRA) and the Whittle's Estimator (WE) through a series of simulation studies have been conduc...

A comparison of Estimation Techniques for electrocardiogram classification

Electrocardiograms are one of the most important bio-signals used to diagnose heart conditions su... more Electrocardiograms are one of the most important bio-signals used to diagnose heart conditions such as arrhythmia. In this paper we interpret the rhythm and identify the potential disruptions using computer assisted methods of interpretation. Moreover, Hannan-Rissanen algorithm (HRA) and Maximum Likelihood (MLE) are used to estimate the coefficients of the Generalized Autoregressive Moving Average (GARMA). We classified ECG signals into 6 classes (normal and 5 types of arrhythmia) using neural networks approach. We also compared the results of HRA and MLE estimations methods and found that HRA performs better than MLE. The overall detection accuracy using HRA is 81.7% and 67.3% for MLE.