Browsing by Author "Dissanayaka, D. M. P. V."

Now showing 1 - 7 of 7

Analysis of the interrelationship between weather parameters in Colombo area
(Faculty of Science, University of Kelaniya Sri Lanka, 2023) Samarasinghe, D. G. S. P.; Hewaarachchi, A. P.; Dissanayaka, D. M. P. V.
Colombo serves as the administrative, legal, and primary urban hub of Sri Lanka in terms of population. Its metropolitan vicinity houses around 5.6 million residents, with 752,993 within the Municipality. The city of Colombo is renowned for its tropical climate, characterized by consistently high average temperatures, substantial annual rainfall, and other meteorological factors. This study attempted to investigate the variations in the weather parameters and to model the interdependencies among those variables. The correlations of five weather parameters between January 2007 and May 2022 were analysed based on the monthly data collected from the Department of Meteorology in Colombo area. Rainfall, Minimum Relative Humidity, Maximum Relative Humidity, Wind Speed, and Temperature were considered in this study. Investigation of the correlation among various weather parameters holds paramount importance in understanding the intricate dynamics of Colombo's climate. A seasonal pattern was found in Minimum Relative Humidity, and high fluctuations were observed in Wind Speed and Rainfall out of the five weather parameters under study. Further, the study concluded that there is a moderate positive correlation (r = 0.5) between Rainfall and Minimum Relative Humidity and as well as between Rainfall and Maximum Relative Humidity (r = 0.7). A moderate negative relationship (r = -0.5) between Wind Speed and Maximum Relative Humidity is also found in this case study. In this research, vector autoregressive (VAR) models were employed to capture the relationships among weather parameters which indicated the presence of Granger Causality. According to the Granger Causality test it was found that Minimum Relative Humidity can be used in predicting the other four parameters i.e., Rainfall, Maximum Relative Humidity, Wind Speed, and Temperature. Maximum Relative Humidity can be used in predicting the Minimum Relative Humidity, Wind Speed, and Temperature. Also, Wind Speed can be used in predicting all other four weather parameters concerned in this study. It has been found that monthly average Temperature has the potential to serve as a predictive factor for all three of the weather parameters except Wind Speed under consideration in this investigation.
An analysis of the technical efficiency of protected agriculture in dry zone area of Sri Lanka
(Faculty of Science, University of Kelaniya, Sri Lanka, 2021) Wickramasinghe, H. M. K.; Dissanayaka, D. M. P. V.
Efficiency is an important factor in productivity growth. It helps in utilizing scarce resources in an effective manner to derive efficient consumption. Identifying the determinants of efficiency levels is a major task in efficiency analysis. Vegetable Production is the most essential sector of the Sri Lankan economy, but recently the seasonality of production supplemented by the climate change effects has posed a major threat to maintaining a constant year-round vegetable production. Protected Agriculture (PA) is identified as one of the best adaptation methods to increase the productivity of horticultural crops, particularly vegetables. The resources that affect the inefficiency in protected agriculture has contributed to such variations in the productivity across the dry zone area. Technical Efficiency (TE) is used as an indicator to determine whether the output reached at its optimal level in production. Both technical and socio-economic factors may result the inefficiency. On the other hand, if it is possible to identify the characteristics of efficient farmers and inefficient farmers, it would be helpful to improve the efficiency of protected agriculture production. Therefore, the attempt of this thesis is to first estimate the technical efficiency in PA based on the factors which derive efficiency of Protected farming and thereby, investigate the influence of socioeconomic factors of farmers on the efficiency in the dry zone. In comparison to the wet and intermediate zones, the dry zone has the highest vegetable production area in the country with sample resources and there are more protected agriculture farmers in the dry zone area. Therefore, the research study was done using all PA farmers in dry zone area. The data set was gathered from the HARTI and consists of 70 farmers' data from the year 2017, taking into account the population of the dry zone area's geographic location. The stochastic frontier approach has been used to generate technical efficiency estimates using Frontier 4.1 by Coeili (1994). The results of this thesis show that the estimated mean technical efficiency of PA production is 49.64 %. Therefore, there is a 50.36 % scope for increasing technical efficiency in PA by using the present technology. The elasticity of inputs is computed from the estimated Cobb- Douglas production function. A production function is a mathematical expression that describes how the quantity of output changes as a function of the inputs utilized in the process. Which concludes, total labor, total Fertilizer, unit price, access to extension service, and initial cost are statistically significant at 5 percent. This implies that the variables of significance remain an important contribution to the determination of technical efficiency in protected farming in the dry zone areas. The analysis reveals that the sum of the partial output elasticity with respect to all inputs is 159, which indicates an increasing return to scale in PA production. Future, it has been found that age, gender, education, and farm size are statistically significant determinants of technical efficiency. The result of the study indicates that the current use of all production inputs is not at the optimum level. Thus, the study provides guidance in increasing the technical efficiency in Protected agriculture in the dry zone in Sri Lanka.
An anthropometric index to estimate the obesity
(Faculty of Science, University of Kelaniya, Sri Lanka, 2021) Tissera, H. A. N. S.; Munasinghe, M. A. H. C.; Dissanayaka, D. M. P. V.
Obesity is a disease that occurs when the percentage of body fat has a negative effect on a person’s health. As for the World Health Organization's definition, obesity is defined as the condition of the body when the body fat is greater than 25% and 35% for men and women respectively. Obesity is a crucial point to discuss as it has been considered a major nutritional health problem in developed and developing countries. Therefore, many indices have been developed to estimate body fat using various measurements of the body. The objective of this study was to develop a simple anthropometric linear equation (index) that is more accurate than the Body Mass Index (BMI) and other indices which currently use to estimate whole-body fat percentage among individuals. Developing a new index to measure body fat is significant as the current indices fail to measure body fat accurately in some exceptional cases like professional athletes. As for an example, the BMI also does not capture information on the mass of fat in different body parts. Hence developing a new index to measure the body fat level is essential. This study used secondary data from the National Health and Nutrition Examination Survey (NHANES) in 2017- 2018. Missing values were imputed by using the multiple imputation techniques. Initially a descriptive analysis was performed to analyze the composition of the sample. It was discovered that the mean fat percentage was 35.416 with a standard deviation of 7.109 and 24.461 with a standard deviation of 7.964 in girls and boys of age 15 to 19 years, respectively. Total fat percentage was considered as the response variable. Simple linear regression models were fitted to find the most correlated variables with the total body fat level. 15 anthropometric indices were generated using transformations on explanatory variables. The best-fitted equation was selected by considering the High Correlation with body fat, Minimum Akaike Information Criterion (AIC), and Highest R2 value. The accuracy of the index was tested using the test dataset and compared with the accuracy of the current indices. It was revealed that this index measures body fat more accurately than the Body Mass Index (BMI) and Waist-Height Ratio (WHR) with an accuracy of 76.8%. Waist Circumference, Hip Circumference and height measurements used to develop new index. Then the selected variables were used with the age category and gender as explanatory variables to perform a multiple linear regression model to find the determinants of the body fat level. As a further study, the developed index can be improved by adjusting for gender-wise and age-wise to obtain more accurate results.
A comparison of distance-based and model-based clustering methods
(Faculty of Science, University of Kelaniya Sri Lanka, 2023) Nadeekantha, H. A. D. D.; Kavinga, H. W. B.; Gunawardana, A.; Dissanayaka, D. M. P. V.
Most of the statistical techniques assume the homogeneity of the sample data. However, not all the time, real-world samples are homogeneous. The existence of subgroups within a population leads to the non-homogeneity of the sample. In this case, it is not accurate to model the population using a single probability distribution. Hence it is essential to check the homogeneity of the sample. Clustering, an unsupervised learning technique, is being used to discover a population's subgroups and group each observation into a specific cluster. Mainly, clustering algorithms can be divided into two groups, namely model-based and distance-based algorithms. Model-based algorithms assume a probability distribution for clustering, while distance-based algorithms use a distance metric to classify observations into clusters. In the literature, it was suggested that the model-based clustering methods perform better than the distance-based methods using summary statistics and visualizations. In this study, an inference-based procedure has been used to assess the above claim. To compare the performances of model-based and distance-based algorithms, an extensive simulation study was conducted. In the simulation study, two univariate Gaussian mixtures with different parameter settings (mean, standard deviation, and sample size) were combined to generate a non-homogeneous sample. Then, model-based and distance-based algorithms were applied to the same simulated datasets with different cluster structures, knowing the actual cluster memberships. Further, the effect of bimodality conditions of Gaussian mixtures on both clustering methods was checked. To assess the performance of the two methods, identifying the correct number of clusters, Cluster Identification Ability (CIA), and categorizing the observations into the correct cluster memberships (clustering accuracy) were computed. CIA was computed using the percentage of iterations that identified the correct number of clusters, and clustering accuracy was measured using the Adjusted Rand Index (ARI). For most of the simulation settings, both methods required a sample size of less than 200 to achieve high clustering accuracy (approximately mean ARI value of 0.8). For example, a simulation setting with a mean difference of 3.1 and a standard deviation of 0.5 required sample sizes 20 and 10 for the model-based and distance-based methods, respectively. These minimum sample sizes vary depending on the method's high clustering accuracy, and in some cases, those are approximately the same. The inference-based study which is performed using the paired Wilcoxon signed-rank test indicated that the claim “model-based method outperforms distance-based method, or both performs similarly” is valid 82.7% of the time at a 5% level of significance. In conclusion, the CIA and clustering ability of the model-based method increased with the increment of sample size when the bimodality conditions were satisfied by the mixture. For the distance-based method, both abilities decreased as the sample size increased when the bimodality conditions were not satisfied by the sample.
Detecting abrupt changes in thermal electricity production data in Sri Lanka
(Faculty of Science, University of Kelaniya Sri Lanka, 2022) Sumathipala, P. L. N. S.; Hewaarachchi, A. P.; Dissanayaka, D. M. P. V.
A changepoint or an abrupt change is a distributional change in a time series data structure. Over the past years, many studies have been conducted to search these changepoints and many researchers proposed several multiple changepoint detection methods. One such search method is the Pruned Exact Linear Time (PELT) method, which is exact and under mild conditions, has a computational cost which is linear in the number of data points. This method is a more accurate and faster method to detect multiple changepoints. The objectives of this study are to detect abrupt changes in thermal electricity data in Sri Lanka and predict thermal electricity production accurately. Since undetected changepoints may cause incorrect modelling or prediction, the accurate analysis of electricity data is vital. In this study, electricity production (Hydro, Thermal oil and coal, and wind) by Ceylon Electricity Board (CEB) in Sri Lanka for the period 2000 to 2019 was used to find abrupt changes. The PELT method is used to detect these changepoints and their location in the variance of electricity data. First, the total electricity production of oil and thermal data were used and a changepoint was found in April 2011. This is a documented changepoint since, according to CEB Annual Report 2011, 1487 GWh of thermal (oil) power was added to the system during 2011, which was a significant change. Moreover, two models, for the periods 2000 to 2019 and 2011 to 2019 (after the detected changepoint) were fitted for forecasting the production. Root Mean Square Error (RMSE) and Mean Absolute Percentage Error (MAPE) were calculated to compare the forecasting accuracy of these models. The first model (ARIMA (2,1,3)), which does not consider the changepoint structure, results in RMSE and MAPE values of 0.911 and 6.009, respectively, for the period 2000 to 2018 for the thermal electricity data. For the second model (ARIMA (1,0,3)), RMSE and MAPE were 0.244 and 3.267, respectively, for the period 2011 to 2018. It can be seen that the models fitted by considering changepoints give more accurate results for forecasting electricity production.
Prevalence of known diabetes in Sri Lanka: results from the Sri Lanka demographic and health survey 2016
(Faculty of Graduate Studies, University of Kelaniya Sri Lanka, 2022) Munasinghe, M. A. H. C.; Nugawela, M.; Jayasundara, D. D. M.; Dissanayaka, D. M. P. V.; Sivaprasad, S.
Diabetes is a major global public health burden. According to International Federation of Diabetes (IDF), Sri Lanka shows an increasing prevalence of diabetes. There is a paucity of contemporary data on the prevalence of diabetes in Sri Lanka. Therefore, this study was conducted to estimate the national and provincial level prevalence of diabetes and establish the demographic risk factors of diabetes in Sri Lanka. We used data from the Sri Lanka Demographic and Health Survey (SLDHS) 2016 conducted by the Department of Census and Statistics Sri Lanka. A total of 106,466 individuals were included in this survey. From the survey data, a total of 71066 individuals aged 20 years and older were identified from all the nine provinces and the diabetes status in the questionnaire was used to define people with known diabetes. Age, gender, ethnicity, religion, education level, smoking history, marital status, urban or rural location, province of residence was included as potential exposures. The outcome was defined as self-reported prevalence of diabetes status. Age adjusted prevalence values were obtained by multiplying the crude age-specific prevalence of diabetes by age-specific weights. Weights were calculated using the Census of Population and Housing (CPH) 2012 data. Multivariable logistic regression was fitted, and Odds Ratios (ORs) were derived to examine the relationship between the covariates and outcome (diabetes status). The age adjusted national prevalence of diabetes is 10.6%. The prevalence of diabetes was higher in women than in men. Provinces with higher GDP (Gross Domestic Product), seemed to have a higher prevalence of diabetes. Prevalence of diabetes was higher in urban residents (14.39%: 95% CI: 13.72% -15.06%) compared to their counterparts in rural (11.38%: 95% CI: 11.10%-11.66%) and estate areas (9.15%: 95% CI: 8.25%-10.04%). The multivariable logistic regression analysis showed that age, urban area, moors, females, province, and high level of education as independent risk factors for diabetes. Moors had 43% increased odds of diabetes compared with Sinhalese (OR:1.43, 95% CI 1.30,1.58). Compared to residing in Rural areas, Urban sector had 19% increased odds of diabetes (OR:1.19, 95% CI (1.11, 1.28)). Females’ risk of getting diabetes was 72% higher than males (OR:1.72, 95% CI 1.62,1.82). Individuals who had a high level of education had 10% of increased risk of getting diabetes (OR:1.1,95% CI 1.04,1.17) than others. People living in Western province, were 64% more likely to have diabetes compared to other provinces. Smoking status of the individuals was not related to diabetes in this analysis. The findings clearly show that known diabetes prevalence in Sri Lanka varies between provinces, with most urban and economically developed regions showing a high prevalence of known diabetes. Given the limited resources available in the health system in Sri Lanka, this study highlights how the population can be stratified for efficient optimization of diabetes care in the country.
A quantitative characterization of sector-wise performance interdependencies in stock market using changepoint and performance-induced distant clustering
(Faculty of Science, University of Kelaniya Sri Lanka, 2023) Kavishka, R.T.; Liyanage, U. P.; Hewaarachchi, A. P.; Dissanayaka, D. M. P. V.
The Colombo Stock Exchange (CSE) is the Sri Lankan marketplace for companies to trade their stocks to the public. There are 19 sectors in the CSE before the Global Industry Classification Standard (GICS) classification. The stock market has become a significant icon in most of the country’s economy. Due to the enhancement of the data science discipline, quantitative research on the stock market has gained popularity among scholars in the recent past. Most of the studies were conducted to predict the value of a stock and its volatility. However, this study explored potential performance dependencies among the 19 industrial sectors registered at the CSE. In known literature, this scope has not been addressed quantitatively. The sectorial All Share Price Index (ASPI) is used to characterize the sector interdependencies, as the volatility of ASPI implies the sector performance at a given time or a short period. Because diverse sector movements can offset each other, leading to a stable index, while extreme sector-specific events or trends can result in increased index volatility. The ASPI indices published by CSE from 2005 to 2019 were considered in the analysis. The persistency of ASPI volatility in a compact interval indicates the consistency of the performance in each sector. Thus, the comparison of volatility changes and their changing time, i.e., changepoint analysis, describes the changes in the sectorial performances. Consequently, the interdependencies among the sector-wise performances can be recognized by the emerging patterns of the changepoints, i.e., the clustered behaviour of the changepoints. Through this approach, the investigation seeks to identify significant transitions or shifts in the behaviour of each sector. Non-parametric methods were employed in the identification of the changepoints of the ASPI series. The standard clustering approaches could not be utilized in grouping the changepoints as the clustering metrics defined by the variation of performances are interconnected. Thus, a new clustering approach was developed using a cluster performance-induced distant measure defined based on a reference industrial sector. This analysis resulted in interdependency among the industrial sectors. Further, variation patterns among the changepoints were identified using interval scaling, and dependent industrial sectors were identified with the help of the performance-induced distant clustering approach. For example, the Bank Finance & Insurance, Telecommunication, and Trading sectors exhibited strong interdependencies. Also, the Construction & Engineering, Oil palms, and Hotel & Travels sectors exhibited strong interdependencies. So far, in qualitative relations, these interdependencies were merely recognized by the gut feelings of financial analysts. Nevertheless, this study provides a clear quantitative characterization of performance dependencies, and thus, the findings are crucial for determining investment strategies and minimizing risk in stock exchanges.