Browsing by Author "Jayasundara, D.D.M."

Now showing 1 - 16 of 16

An Application of 5-fold Cross Validation on a Binary Logistic Regression Model
(2016) Attanayake, A.M.C.H; Jayasundara, D.D.M.; Peiris, T.S.G.
Abstract Internal validation techniques can be used to check the predictive ability of the developed models. The most common internal validation techniques are split sample methods, cross validation methods and bootstrapping methods. The split sample methods are inefficient with the small size of data sets. The bootstrapping methods are efficient with the knowledge of computer programming languages. The cross validation methods are not very popular in practice. Therefore, in this study 5-fold cross validation method of cross validation techniques is applied to validate the predictive ability of a binary logistic regression model. The binary logistic regression model was fitted on a data set of UCI machine learning repository. Results of the cross validation reveal that low value of optimism and high value of c-statistic in the fitted regression model indicate an acceptable discrimination power of the developed model.
Forecasting Exchange Rates using Time Series and Neural Network Approaches
(Proceedings of the 69th Annual Session of the Sri Lanka Association for Advancement of Science -SLASS, 2013) Nanayakkara, K.A.D.S.A.; Chandrasekara, N.V.; Jayasundara, D.D.M.
Forecasting exchange rates using time series and neural network approaches
(Sri Lanka Association for the Advancement of Science, 2013) Nanayakkara, K.A.D.S.A.; Chandrasekara, N.V.; Jayasundara, D.D.M.
Identifying Factors affecting the severity of Asthma
(European International Journal of Science and Technology, 2013) Peramuna, P.R.N.L.; Chandrasekara, N.V.; Jayasundara, D.D.M.
Identifying factors associated with severity of asthma in Kegalle District
(Sri Lanka Association for the Advancement of Science, 2013) Peramuna, P.R.N.L.; Chandrasekara, N.V.; Jayasundara, D.D.M.
The impact of mothers’ migration for work abroad on children’s education
(Sri Lanka Association for the Advancement of Science, 2013) Dissanayake, P.L.; Chandrasekara, N.V.; Jayasundara, D.D.M.
Impact of Mothers' Migration on Children's Education
(Proceedings of the 69th Annual Session of the Sri Lanka Association for Advancement of Science -SLASS, 2013) Dissanayake, P.L.; Chandrasekara, N.V.; Jayasundara, D.D.M.
Level of stress among Advanced Level students: A case study
(University of Kelaniya, 2013) Arachchi, A.P.K.; Jayasundara, D.D.M.
The purpose of this study is to determine the factors that affect the stress level of A/L students and determine the relationships between stress level and those factors. Stress affects people’s behaviors, communications, and efficiency. Stress is not only a factor in workplaces; it is also a common factor in educational environments experienced by students. The participants of this study are 300 Advanced Level (2013) students of secondary schools in the Minuwangoda Zone. The reason for choosing Minuwangoda Zone is mainly because it has a high success rate among the schools in the Gampaha District, in terms of Advanced Level results. Leading A’Level schools are found in this zone. The main objective of the study is to describe the effect of study habits, parental attachment, demographics and attitudes of students on the level of stress. Stratified random sampling was used to select a proper sample size for the analysis. The population of students was stratified under three factors: School Category, Gender, and Subject Stream. A questionnaire was designed to collect relevant information about gender, subject stream, attitude towards subjects, physical health, study habits, parental attachment and the response level of stress. Data was analysed using descriptive statistics, graphical summaries, non-parametric tests and finally a logistic regression model. The analysis showed that factors like concentration, lack of time at exams, trust of parents, sleepiness while studying, number of tuition classes, and being allocated space for studying, are contributory factors to the level of stress of students.
Linear programming approach to assess an optimal cultivation plan: A case study
(Faculty of Science, University of Kelaniya, Sri Lanka, 2020) Hakmanage, N.M.; Chandrasekara, N.V.; Jayasundara, D.D.M.
An optimal cultivation plan refers to the procedure or action of making the best or most effective use of resources for cultivation in a sustainable manner while maximizing net return. Reaching an efficient cultivation plan and utilization of resources and requirements is often a challenging problem. To utilize resources and requirements such as water, land, manpower, fertilizers and seeds, optimization techniques are used. The objective of this research is to maximize the net return of the cultivation using linear programming technique and allocate the arable land optimally. Linear programming is the most convenient and effective tool to handle the objective function with many constraints. This study was carried out in a rural village located in Dompe divisional secretariat in Gampaha district using 150 farming lands, to determine the land resource allocation for twelve selected crops: bitter gourd, lady's fingers, manioc, potatoes, rambutan, banana, pineapple, beetle, rice, coconut, tea and pepper. The linear programming model is formulated for the optimal land resource allocation of 4477.2 perches. The maximum net return projected by the proposed model is Rs 6,370,512.00 for cultivation seasons. The proposed solution is a 34.96% increase in profit as compared to the actual profit obtained from the cultivations. Crops like rambutan, rice, manioc and pineapple which provides a higher return should be developed and cultivation extended under the supervision of the agricultural expertise or officers. The model suggests that some crops such as lady’s fingers, potatoes, banana and coconut may not be providing comparable returns versus the other selected crops. The results reveal that linear programming approach will significantly improve the net benefits with optimal crop area allocation. The limitation of this study is that it is considered the soil condition is the same for all crops in the study area. Advanced operations research techniques like multi objective nonlinear programming models will be employed for this study in future work.
Modeling and Forecasting Selected Climatic Factors Influenced on Sustainable Cultivation Plan: A Case Study for Dompe-Gampaha District
(International Postgraduate Research Conference 2019, Faculty of Graduate Studies, University of Kelaniya, Sri Lanka, 2019) Hakmanage, N.M.; Chandrasekara, N.V.; Jayasundara, D.D.M.
The agriculture is the back born of economy of the most Asian countries. Although the country is moving towards industrialization, the agricultural sector still continues to be an important sector in the economy in Sri Lanka. Cultivation is the predominant sector of the agriculture. Lack of sufficient amount of water is the main limitation factor for cultivation while flood/ deluge is causing the waste of harvest. The main water source for cultivation in Sri Lanka is rainfall. Moreover, for each crop due to its peculiarities and mainly owing to its geographical origin, there exist specific temperature limits within which these plants are able to grow and reproduce. Hence rainfall and temperature are imperative factors influenced on cultivation. More accurate forecasting of monthly rainfall and temperature is significantly important in irrigation schedule, water resources management, crop pattern design and designing of harvesting amount. The main objective of this study is to build suitable forecasting models for two climatic factors: Temperature and Rainfall which affect sustainable cultivation plan. Monthly data of rainfall and temperature from 2009 to 2019 of Dompe-Gampaha district was considered for the study. First 80% of data was used to formulate the models and the rest 20% data was used to validate the models. The paper introduces two fundamentally different approaches for designing a model, the statistical method based on seasonal autoregressive integrated moving average (SARIMA) and decomposed ARIMA model. Mean absolute percentage error (MAPE), Root mean squared error (RMSE) was used to evaluate the performance of fitted models. Among the fitted ten models, SARIMA(0,0,0)(1,0,1)12 was identified as the better model to forecast rainfall based on minimum Akaike information criterion (AIC) where MAPE and RMSE are 48.57% and 5.1339 respectively. Although Box Ljung lack of fit test prove that this model is suitable model, the errors are extremely high. Then decompose ARIMA model was used by calculating seasonal and trend component using SARIMA(0,0,0)(0,1,0)12 and linear regression (Trend=14.33321–0.04884*time) models respectively. Summation of forecasted values of these two models is the forecasted value of decompose ARIMA model and it exhibits MAPE which is 20% lower than the SARIMA(0,0,0)(1,0,1)12 model. Therefore, fitted decomposed ARIMA model can be recommend as a better model to forecast rainfall of Dompe-Gampaha district. Similar approach was carried out to find a suitable model to forecast temperature. SARIMA(1,0,0)(2,0,1)12 was the most accurate model to forecast temperature with minimum AIC value. MAPE and RMSE of this model was 1.3938% and 0.4695 respectively. Lack of fit test and errors provide evidence to say that the fitted SARIMA(1,0,0)(2,0,1)12 is suitable to forecast temperature in the study area. The forecasted values of rainfall and temperature can be used when developing sustainable cultivation plan in Dompe-Gampaha district which leads to development of agricultural sector of the country
Modeling inflation in Sri Lanka using time series and econometric approaches
(University of Kelaniya, 2013) Udani, S.A.L.M.; Jayasundara, D.D.M.
Inflation is the most important macro economic variable often used to measure the economic growth of a country. It is the rise of the general level of prices of goods and services in an economy over a period of time. In Sri Lanka, it is expressed as the percentage change of the Colombo Consumer Price Index (CCPI). Modeling can be done either using time series approach or econometric approach. Time series approach predicts inflation through modeling the past annual inflation figures only; while econometric approach considers the impact of other economic factors on annual inflation rates. In time series approach an ARIMA(1,1,1) model was fitted to the annual inflation rates of 1953 to 2001 using MINITAB since it is the best model to forecast future inflation rates. The fitted values and the actual figures were approximately similar and the actual values were within the upper limits and lower limits. Econometric approach was conducted by fitting a cointegration equation and vector error correction model (VECM) for the annual inflation rates of 1977 to 2004 using EViews. Since the data series were cointegrated, the ordinary least square (OLS) model was inappropriate. The cointegration equation revealed long run relationship of inflation. Money supply and exchange rate played a significant role of the long run relationship. A vector error correction model implied a short run relationship of inflation. The R-squared value of VECM model showed that real gross domestic product, money supply and exchange rate explain 86.83% of the variance of inflation. It implied that the impact of percentage changes at past years of money supply, exchange rate and real gross domestic product affected the current level of inflation, highly. Both exchange rate and money supply have significant impacts on long run and short run relationships of inflation.
Optimal assignment of unusable/ waste lands effectively using improved fuzzy assignment technique
(Faculty of Graduate Studies - University of Kelaniya, Sri Lanka, 2021) Hakmanage, N.M.; Jayasundara, D.D.M.; Chandrasekara, N.V.
Land resources are valuable for humans not only live but also conduct all of their economic activities on it. Allocation of land uses in a critical and optimal manner will pave the way for determining policies for the optimal utilization of land in a sustainable manner for the future, focusing on the uncertain conditions of each allocation. The objective of this study is to identify and propose effective allocations to abandoned lands such as unusable, waste and uncultivable lands using optimal land assignment plan. Fuzzy assignment technique accesses to explore how uncertainty in suitability index and the condition of the land will affect to optimal land allocation with the minimum allocation cost in this study. A major land-use classification system in Sri Lanka contains multiple levels of classification. Among them, land use categories regarded to the study area (farming village which has six unusable lands in Dompe divisional secretariat, Gampaha District) are selected as follows: Agriculture, Habitable or settled lands (Urban or rural areas), Forests, Wildlife, Reserves & Catchments areas, Underutilized Lands, Reservations (Reservoirs, Streams, & Irrigation Channels) and Barren lands. Major properties of the land were identified as land area: vaguely defined categories measured in square meters; Ownership: three possible sectors according to the ownership of the land as Private, Public and Other; Condition: discretional parameter that is vaguely defined with three possible values: bad (0), average (0.5) and good (1) and the Facilities: four different categories (power (P), water (W), communication (C), transportation (T)). Subsequently, the properties of each land and all possible demands were identified and a suitability index was developed using those vague parameters for each assignment of lands. With the aid of the Center of Gravity (COG) method, fuzzy values were converted to their crisp equivalents. Then the cost of assignment of each land for the aforementioned purposes, were considered using with linear, triangular, and trapezoidal fuzzy membership intervals. Thereafter, Robust ranking technique was applied to calculate the numerical values for the interval and obtain the product of suitability index and cost of allocation. Finally, using the Hungarian assignment algorithm, each land was assigned optimally for its effective purposes. The linear, triangular, and trapezoidal membership degrees, the minimum cost was obtained from the trapezoidal membership degree, that is 15% lower than the linear membership degree. Therefore, study proceeds with the trapezoidal membership degree. Using hypothetical assignment costs, six lands in the study area were assigned optimally for agriculture, habitable or settled lands, forests, wildlife, reserves and catchments areas, underutilized lands, reservations, and barren lands. This will be a great social and environmental service as it will involve the re-usage of the lands that are currently abandoned. Furthermore, the findings of this study can be extended nationally to save and maintain the land resource in an optimal manner.
Reduction of experimental error in coconut research by choosing proper data analysis techniques
(Faculty of Science, University of Kelaniya, Sri Lanka, 2020) Fernando, W.H.H.; Waidyarathn, K.P.; Jayasundara, D.D.M.
Accurate data analysis techniques are essential in field experiments to correctly understand the influence of independent variables on the dependent variable/s. This study compares different statistical techniques used in analyzing longitudinal data (nut yield data collected in multiple years) of a coconut field experiment. Longitudinal studies are necessary for coconut research due to its perennial nature. However, these experiments often have high variability due to the heterogeneity nature of coconut palms, where individual palms display inconsistent temporal behaviour. High variation among the individuals in similarly treated plots makes treatment mean sensitive to those fluctuations ultimately masking the true treatment effect. Even careful planning of the experiment cannot ensure the total elimination of this component. The study highlights the ways in which how this unaccountable variability should be handled to obtain a precise research output. There are many types of statistical techniques used in analyzing data from different experimental designs to achieve optimal research outcomes. This study compares different statistical techniques applied to a randomized complete block design (RCBD), the most frequently used experimental design in coconut research, using a long term coconut fertilizer study. The example illustrates the appropriate types of analyses to meet the precise analysis output by evaluating the model residuals and the Coefficient of Variations (CV). CV, the ratio of the standard deviation to the mean (total average of the design), is a measure of relative variability. In particular ANOVA, Mean Square Error of ANOVA can be used as the standard deviation of the design because Standard Error (SE) of a statistic (usually an estimate of a parameter) represents the standard deviation of its sampling distribution. In this study, Repeated Measures Analysis of Variance (RMANOVA) was used as the classical method. Improved methods used were Linear mixed model and Multivariate Analysis of Variance(MANOVA) with two principal components (representing ≥ 78% variation of the data) as dependent variables. Adequacy of all methods was accepted after checking normality with the Shapiro-Wilk test, homogeneity of variance with Levene’s test, and independence of residuals with the Box-Pierce test. CV resulted from RMANOVA applied on RCBD was 39.2%, while it was 16.51% from the Liner mixed model. The lowest CV (10.04%) resulted from MANOVA with two principal components indicates that it can be more efficiently used to analyze long term experiments of coconuts. The consistency of the results should be studied further with a few more similar kinds of data sets. In addition to the above statistical analysis techniques, Bayesian inference methods will be studied for further improvements in the results.
Statistical and Mathematical Models for a Sustainable Cultivation Plan - A Brief Review of the Literature
(International Conference on Applied Social Statistics (ICASS) - 2019, Department of Social Statistics, Faculty of Social Sciences, University of Kelaniya, Sri Lanka, 2019) Hakmanage, N.M.; Chandrasekara, N.V.; Jayasundara, D.D.M.
Due to rapid increase of population, demand for food is increasing. If the agricultural sector fails to supply and meet the rising demand of foods, it will affect the economy. This however requires finding viable solution that is balanced supply demand food chain. To fulfil subsistence food needs there should be a stationary state with proper cultivation plans in agricultural sector in a country. This study presents a review of the literature published between 1998 and 2019 on cultivation plans in the areas of statistics and/ or mathematics which have considered major influential factors for cultivation. This study aims at reviewing the most appropriate sub sections: arable land selection, cultivating methodologies and climatic factors effect on cultivation to build an optimal cultivation plan. Review was conducted using separate articles as searching strategy was failed to identify published articles which studied for these three aspects together. Hence the significance of this study is to discuss how to apply statistical and/ or mathematical models which are used to implement the cultivation plan including all influential factors together
The impact of mothers' migration for work abroad on children's education
(Proceedings of the 69th Annual Sessions, 2013) Dissanayake, P.L.; Chandrasekara, N.V.; Jayasundara, D.D.M.
Time Fluctuation Models to Forecast Tea Production, Prices and Exports in Sri Lanka
(University of Kelaniya, 2012) Aponsu, G.M.L.M.; Jayasundara, D.D.M.
The production and consumption of tea worldwide have increased over the past decade, and this increase is expected to continue. The tea industry is a key component in the Sri Lankan economy, and is a leading foreign income generator. Sri Lanka is the world's fourth largest producer of tea. The industry mainly consists of tea production, tea export and tea auctions. Being a crop which contributes greatly to the Sri Lankan economy, it is very important to be aware of the future fluctuations in tea production, prices and exports. The objective of this study is to identify predictive models to forecast monthly tea production, prices and exports. Based on the time series analysis, Seasonal Auto Regressive Integrated Moving Average (SARIMA) models were fitted to forecast the monthly tea production and tea exports. The study was undertaken by using the data from the Sri Lanka Tea Board. Monthly black tea production and price data were considered from 1988 to 2009, and tea export data for six tea categories namely Bulk, Green, Instant, Packeted, Other tea and Tea bags, were considered from 1996 to 2010. There was a little upward trend and a seasonal pattern in both production and export data. Since the series is non-stationary, seasonal differencing procedure was required and the 1st seasonal differenced series was generated by a non-stationary process. With the use of calculated parameter estimates, Box-Pierce (Ljung-Box) Chi-Square statistic and residual analysis, it was found that the most suitable model to forecast the monthly black tea production is SARIMA(3,0,3)(0,1,1)6 and the best model to forecast total tea export is SARIMA(1,0,2)(0,1,1)12. Monthly average black tea prices at Colombo auctions were forecast using polynomial regression model. Among linear, quadratic and cubic regression models, cubic regression model was the most suitable model with the R2 value of 91.3%. The fitted regression equation is, Price = 34.36 + 0.9943 (Time Index) - 0.006900 (Time Index)2 + 0.000029 (Time Index). Time Index is the notation of time between the considered time period in months and it explains the tea prices with an accuracy of 91.3%.The identified forecasting models will help to study future fluctuations in the tea industry.