Search

Deteriorating air quality and uncontrolled air pollution can affect climate change as well as human health. The Air Pollutant Index (API) has been used as an indicator to indicate the status of air quality. The objective of this study is to determine the level of air pollution based on the calculation of API as well as to identify clusters for air quality monitoring stations based on the concentration of air pollutants and API values. This study involved 10 air quality monitoring stations from the urban, sub-urban and industrial areas. The data used in this study are the daily concentration of air pollutants namely carbon monoxide (CO), ozone (O3), fine particles (PM10), sulfur dioxide (SO2) and nitrogen dioxide (NO2) from January until December 2011. Median of the pollutant concentrations were used to represent the monthly API for each station. Urban areas had the highest percentage of moderate air quality levels compared to suburban and industrial areas. O3 was found to be the dominant pollutant in urban areas while PM10 is the dominant pollutant in suburban and industrial areas. Fuzzy k-medoid clustering (FKM) is one of the fuzzy clustering algorithms that has a cluster center known as the medoid and can produce soft clusters. The air quality monitoring stations are assigned to the clusters according to the membership degrees ranging in the interval [0,1] and the data points of the station can belong to more than one cluster at the same time with different membership degrees. The degree of memberships indicate the level of confidence that the data points are in a particular group. Two clusters were obtained in which cluster 1 is characterized as a cluster with higher pollutant concentrations and API values compared to cluster 2. [ABSTRACT FROM AUTHOR]

Medical images are images of the interior body taken using medical imaging techniques. Most of these images are contaminated with noise caused by transmission limitations or storage devices during the process of image collection from MRI, CT scan, or biopsy procedures. To date, many researchers have found that image denoising is a difficult task because they must strike a balance between noise reduction and preserving sharp image edges. Instead of conventional denoising methods, this paper studies the application of Shearlet transform for denoising the medical image. Shearlet transform is able to provide better directional determination during image reconstruction because it has the properties of a multiscale and is able to extract anisotropic features on images. In this study, different thresholding methods such as hard threshold, Otsu threshold and adaptive threshold were considered for denoising the image. In this study, peak signal-to-noise ratio (PSNR), mean square error (MSE) and structural similar index (SSIM) are used for performance evaluation. The experimental results show that the Shearlet transform is suitable to be used for denoising the medical images, where the most suitable thresholding method is the hard threshold, which it effectively reconstructs the images and preserve the edges on medical images compared to the other considered thresholding methods. [ABSTRACT FROM AUTHOR]

Employee turnover is a problem that affects every organization, whether in government or private sector. Employee attrition leads to high costs for any organization, especially in terms of training. In Royal Malaysia Police Force (PDRM), the optional retirement rate is higher compared to those that retire at retirement age. This study will focus on the factors that cause PDRM officers to choose optional retirement using Machine Learning (ML) techniques. In this study, k-prototype cluster analysis, Random Forest, and text analytics were performed for various analysis purposes. The results show that an officer's age is the primary motivator for electing optional retirement over mandatory retirement. Various health issues cause lower productivity and enthusiasm in performing tasks in their work. The work placement, the job ranks, the remainder service year, and the time period of the last promotion is among those identified crucial factors that contributed towards early retirement in PDRM. Since family concerns are frequently cited as a reason for retirees choosing early retirement, the work-life balanced in the police force profession was also noted as another early retirement factor. The findings of this study may assist PDRM in revamping the career in the police force so that the problem of high attrition rate can be curbed and also make the profession more attractive, hence attracting more people to join the police force. [ABSTRACT FROM AUTHOR]

Time series clustering is a data mining technique where it involves grouping of time series data into a few separate clusters. In this paper, a shape-based clustering is applied on time series of PM10 data collected from 60 air quality monitoring stations in Malaysia from 5th July 2017 to 31st December 2018 obtained from Malaysian Department of Environment. k-shape clustering method along with shape-based distance (SBD) were used to cluster the PM10 time series data. Air quality patterns were analysed from the clusters formed. The results show that the clusters formed using the method used are very well separated and mainly influenced by the region and geographical locations instead of the station's categories or man activities at the locations. [ABSTRACT FROM AUTHOR]

Exchange rate instability poses a global economic challenge that impacts international trade and the overall economies of countries, affecting their human development as well. This paper proposes a method that clusters the exchange rate time series of countries and it's compared with the Human Development Index (HDI) categorization. By doing so, a deeper understanding of the relationship between currencies and development status can be gained. In this study, a fuzzy clustering framework was employed to group the time series of exchange rates for specific countries. A total of twenty-five currencies were chosen for analysis during the period from January 01, 2014, to December 31, 2018. Time series clustering was conducted based on the autocorrelation patterns of the time series. The fuzzy c-means approach was used to calculate the similarity between their autocorrelations. The analysis findings indicate the Autocorrelation-FCM approach demonstrated better cluster results compared with the FCM. Moreover, it was observed that the clustering of exchange rates significantly influenced the clustering of HDI. Based on these results, it is recommended that policymakers adopt a diverse range of policies to ensure stability in the exchange rate to foster human development. Such measures hold the potential to bring about significant changes and improvements in monetary policy across different nations. [ABSTRACT FROM AUTHOR]

Analysis on finding patterns of household expenditures including the relationships between various household characteristics and family consumptions is important. This is because the behaviours of household expenditures affect economic developments and policy planning of a country. This paper explores characteristics of households with different levels of household expenditures to determine characteristics of consumers in Malaysia. Drawing from the Household Expenditure Survey (HES) done by the Department of Statistics Malaysia, the relationships between all variables in the household expenditure survey are extracted by using an association rules mining algorithm which is the Apriori algorithm. Association rules mining method is proposed in this study to gather hidden information in the survey since it is able to identify combinations of variables that occur concurrently in large dataset. The Apriori algorithm is chosen because it is simple and easy to execute. The study shows that there are significant relationships between the characteristics of a household with their total expenditures based on chi-squared test. Family size, education background and marital status are all important factors influencing household expenditures. [ABSTRACT FROM AUTHOR]

Degradation models, which are a function of time, are frequently used in reliability analysis. This study uses a linear path degradation model in order to derive the failure time distribution and its percentiles based on parametric and semi-parametric methods. For the parametric technique, which is represented by the maximum likelihood estimator, the random effect parameter is followed the skew normal distribution to estimate the time to failure distribution. While for semi-parametric technique the skew normal distribution beside the classical kernel density function is considered to derive the time-to-failure distribution. The performance of these methods is compared by the criteria bias and mean squared error via a simulation study and unspecified data. In addition, GaAs laser degradation data is provided as an application of real data. Parametric method has a good performance than semi-parametric method under simulated and real data while under unspecified data the semi-parametric method is given a good performance. [ABSTRACT FROM AUTHOR]

Hierarchical clustering is a common method used for monitoring air quality stations. This method groups homogeneous stations and subsequently helps in making decision whether to increase or reduce the number of air monitoring stations in an area. There are numerous dissimilarity measures available to be applied in hierarchical clustering and choosing a suitable dissimilarity measure is crucial. The aim of this study is to compare dissimilarity measures of hierarchical clustering such as Euclidean, Manhattan, Chebyshev, Minkowski and Pearson correlation distance. The hierarchical clustering using the five different dissimilarity measures were performed on the daily concentrations of CO, O2, SO2, O3 and PM10 from January to November 2011 in Peninsular Malaysia. For each cluster obtained in the analysis, the average air quality is observed by calculating the air pollution index (API) of the cluster. The hierarchical clustering found using Euclidean, Manhattan, Chebyshev and Minkowski distance produce similar clusters. Meanwhile, the Pearson correlation distance produce different clusters for almost all of the air pollutants under study. This is because the Pearson correlation distance depend on the strength of relationships between variables while the other dissimilarity measures depend on the standard distance between variables from different stations to form the clusters. Hence, it is sensible to use a dissimilarity measure from the Minkowski family with the Pearson correlation distance to extract information from two different angles in clustering air pollutants. [ABSTRACT FROM AUTHOR]

This study compares three different deep learning approaches for forecasting the air quality time series data. All the approaches were based on the convolutional neural network (CNN), which is one of the deep learning algorithms. The first approach, known as the univariate CNN (UCNN), is based solely on the time series past values. Meanwhile, the second approach, known as the multi-parallel CNN (MPCNN), will use the same variable time series from different observations as the input for the model. The last approach is the multivariate CNN (MCNN), where the forecast will incorporate other variables' time series as the input for the model. The time series data used in this study are the hourly particulate matter, PM10 and PM2.5, concentrations from eight air quality monitoring stations in Peninsular Malaysia. The duration of the time series is from July 2017 until June 2019. The results show that the best air quality predictive model is a model which is built using UCNN since it shows the highest accuracy for both time series PM10 and PM2.5. It is also shown that the forecast will have better accuracy if shorter past values are used as the input for the model. [ABSTRACT FROM AUTHOR]

Medical imaging is an essential element in healthcare, especially for the early diagnosis and treatment of illnesses, including breast cancer. Breast cancer is an illness with a high mortality rate that mostly affects women worldwide. Currently, histopathology image analysis is one of the rising research areas in cancer diagnosis. With digital images, the demand for more accurate and visually pleasing images increases. However, the nature of images is normally degraded by noise which may affect image quality. Nowadays, image denoising has become a common practice in image processing tasks to preserve the quality of images for imaging analysis. Therefore, this study compares the performance of several image denoising methods for histopathology images. The BreakHis dataset is used in this study, which consists of 7909 histopathology images from 82 patients in Parana, Brazil. This study compares several denoising methods such as Gaussian smoothing, anisotropic diffusion, wavelet transform and shearlet transform, when standardized noise was presented on the histopathology images. The performance of denoising methods was examined based on the Peak Signal-to-Noise Ratio (PSNR) and Mean Squared Error (MSE). The results showed that the shearlet provides more efficient performance in denoising the histopathology images than other methods. Shearlets provide an optimally sparse representation of 2D images with multiresolution and multiscale properties. It can be concluded that the shearlet is the best image denoising method for breast cancer histopathology images. The research findings are expected to facilitate the researchers or radiologists in image processing tasks to obtain high-quality images prior to classifying images into benign or malignant tumours. [ABSTRACT FROM AUTHOR]

In this study, the Support Vector Regression (SVR) model was used to forecast air quality time series data which is the particulate matter 10 micrometers or less in diameter (PM10) in Malaysia. Several SVR kernel functions, which are the Linear, Polynomial and Radial Basis Function (RBF) kernels, were considered in this study to determine the most suitable kernel function for forecasting the PM10 time series. The period of the data is from 5th July 2017 to 31st January 2019 consists of five air quality monitoring stations which are Kangar station in Perlis, Tasek Ipoh station in Perak, Shah Alam station in Selangor, Pasir Gudang station in Johor and Kuala Terengganu station in Terengganu. Model performance was compared based on the testing dataset's mean squared error (MSE) values. The results show that the SVR model with Radial Basis Function kernel is more suitable for forecasting the PM10 time series compared to the Linear and Polynomial kernels. [ABSTRACT FROM AUTHOR]

Air is an essential asset in daily life and air pollution is a global problem that concerns society. Apart from obvious factors of air pollution, this study aims to investigate the influence of meteorological parameters such as temperature, humidity, and wind speed on air pollution whether they affect the concentration of pollutants like CO, NO2, SO2, PM10 and O3. Pearson correlation analysis is conducted to determine the significance and strength of relationships between the concentrations of each pollutant with meteorological parameters from January to November 2011 in Malaysia. The results of Pearson correlation analysis show that there is a significant correlation between the concentrations of pollutants with temperature at 5% level of significance. PM10 has a significant negative correlation with humidity while CO, NO2 and O3 negatively correlated at 5% level of significance with wind speed. Then, simple linear, polynomial, exponential, and multiple linear regression models were built to further explore the relationship between meteorological factors and the concentration of air pollutants. The regression models obtained show that CO, NO2, and O3 pollutants are dependent on temperature, humidity, and wind speed where the dependence of these three pollutants on wind velocity is negative while the dependence on temperature and humidity are positive. As for PM10 pollutants, it was found that it has a relationship with both temperature and humidity. Meanwhile, for SO2, it only showed a significant relationship with temperature. Overall, the model and results obtained in this study can be used for air quality prediction and management in Malaysia. [ABSTRACT FROM AUTHOR]

Music is said to play an important role in influencing emotions. This study aims to investigate the influence of music on the emotions and happiness of university students. The study has collected 356 songs which are commonly heard among students from a sample group consisting of 40 university students from the Department of Mathematical Sciences of a public university in Malaysia. After that, audio features for each song are extracted through Spotify Web API. Analysis of audio features has been done on all songs collected and according to the language of the songs. Furthermore, k-means clustering method was used throughout this study to group the students based on the audio features of songs which they are interested in. In addition, in order to measure the level of students' happiness, The Oxford Happiness Questionnaire is answered by each respondent. The results show that students who have no music tendency have the highest happiness level. [ABSTRACT FROM AUTHOR]

Tree-based classification techniques were used to identify the crucial factors influencing the Malaysian family well-being. Data from the nationwide survey, conducted by the National Population and Family Development Board (NPFDB) of Malaysia in 2011 were used for the analysis. Two types of tree-based classification techniques, which are the decision tree and bagging decision tree, were considered for this study. The results from both techniques shown that the satisfaction on family relationships, which is one of the social factors, was the most important factor in determining family well-being. It is also shown that the bagging decision tree performed better compared to the single decision tree. [ABSTRACT FROM AUTHOR]

Twitter is a social media service which enable users to send and also received messages or post known as tweet. In the era of big data, one of the main source of data is from Twitter. However, this types of data usually are unstructured, massive and complicated to be analyzed. In this study, Twitter's data from Malaysian users were collected for one whole week. The objectives are to study the daily rhythm of Twitter usage in Malaysia and also to identify relationships between various types of Twitter users in Malaysia. Since the data are unstructured, pre-processing technique have to be done on the collected data which will enable the data to be analysed. The post-processed data were then analyze by using various statistical analysis tools. From the result of the analysis, we observed some interesting findings which are useful on the describing usage behaviour of Malaysian Twitter users. [ABSTRACT FROM AUTHOR]