Lakshika Nawarathna | University of Peradeniya (original) (raw)

Papers by Lakshika Nawarathna

Transformative Applied Research in Computing, Engineering, Science and Technology, 2025

Identifying the sex of unidentified human skeletons is crucial in bioarchaeology and forensic ant... more Identifying the sex of unidentified human skeletons is crucial in bioarchaeology and forensic anthropology. This study develops a clustering-based approach using 22 measurements from proximal, distal ends, and shafts of both femur bones to determine sex in the current population of Sri Lanka. K-means with hierarchical clustering techniques were evaluated using elbow, silhouette, and gap statistics for K-means, while Ward's method for hierarchical clustering demonstrated higher accuracy in identifying sex. Specifically, Ward's hierarchical clustering method showed an accuracy of 90.6%. Both methods identified two clusters as optimal for sex classification, making them suitable for determining sex in unknown, mutilated, or dismembered skeletal remains.

Measurement error models (MEMs) provide a flexible framework to model the method comparison data ... more Measurement error models (MEMs) provide a flexible framework to model the method comparison data by incorporating measurement errors. However, these models often rely on normality assumptions, which are frequently violated in practice due to skewness and heavy tails. Furthermore, repeated data with measurement errors (MEs) are often observed in medical research, epidemiological studies, economics, and the environment. Thus, this research aims to assess the extent of similarity and agreement between the two methods using the replicated measurement error model (RMEM) under asymmetric and heavy-tailed distributions with a matching degree for true covariate and errors. The expectation-maximization (EM) approach is applied to fit the model. A simulation study is used to test the proposed methodology, demonstrated by evaluating subcutaneous fat data. The Total Deviation Index (TDI) and Concordance Correlation Coefficient (CCC) were used to further assess the agreement between the methods. Our suggested model works well for analyzing replicated method comparison data with measurement errors, skewness, and heavy tails.

Comparing two measurement methods is vital in various fields, such as medical research, epidemiol... more Comparing two measurement methods is vital in various fields, such as medical research, epidemiology, economics, and environmental studies, to determine whether a new measurement method can be used interchangeably with an existing one. Measurement error models (MEMs) are commonly used for this purpose, where the methods have different measuring scales. However, these models often assume normality, which can be problematic when dealing with skewed and heavy-tailed data. To address this issue, we propose the replicated measurement error model (RMEM) under scale mixtures of skew-normal (SMSN) distributions with different levels of skewness and heavy tails of the true covariate and error distributions. Our primary aim is to assess the extent of similarity and agreement between two measurement methods using this model. The expectation conditional maximization (ECM) approach is applied to fit the model. A simulation study is conducted to evaluate the effectiveness and robustness of the proposed methodology and is illustrated by analyzing systolic blood pressure data. The probability of agreement is used further to assess the agreement between the two measurement methods. The findings indicate that the proposed model effectively analyses replicated method comparison data with measurement errors, even when there are outliers, skewness, and heavy-tailedness.

J. Sci. Univ. Kelaniya, 2024

Leptospirosis is a significant health issue in Sri Lanka, but current studies have only examined ... more Leptospirosis is a significant health issue in Sri Lanka, but current studies have only examined some districts using time series analysis. There is no comprehensive leptospirosis prediction model for the entire country. This study aims to develop a model to predict leptospirosis distribution across Sri Lanka. We collected reported leptospirosis cases in all districts from 2011 to 2022 from the Epidemiology Department. We tested the normality of the data using the Shapiro-Wilk test. Also, we evaluated various models, including linear regression, exponential regression, polynomial regression, Autoregressive Integrated Moving Average (ARIMA) time series, generalised additive models (parametric), Cox proportional hazards regression model (semi-parametric), and median-based linear model (nonparametric). The best-fitting model was selected based on the minimum Akaike information criterion (AIC) and Bayesian Information Criteria (BIC), and its accuracy was verified using the Root Mean Square Error (RMSE).

Statistics and Applications {ISSN 2454-7395), 2024

Dental arch dimensions play a crucial role in orthodontic and prosthodontic treatments, defined b... more Dental arch dimensions play a crucial role in orthodontic and prosthodontic treatments, defined by dental arch morphology and widely utilised across various fields. This study focuses on classifying gender based on dental arch dimensions among Sri Lankan individuals. The dataset includes arch dimensions of 573 individuals from different provinces, excluding the Eastern Province. Statistical tests, such as Student's t-test, Kruskal-Walli's test, and Analysis of Variance (ANOVA), were applied to assess the significance of arch dimensions concerning Gender and Ethnicity. Most dimensions showed significance. Various visualizations were employed, and classification methods, including K-nearest neighbours (KNN), Support Vector Machines (SVM), Naïve Bayes (NB) Classifier, Decision tree, and Random Forest (RF) Classifier, were used for gender prediction. Results revealed high accuracy, with the upsampling SVM model achieving the highest accuracy. Sensitivity, specificity, and F1 scores were notable for upsampling SVM models and the rose function RF classifier model, achieving the highest F1 score. In conclusion, upsampling SVM and rose function RF classifier models effectively classify gender based on available dental arch dimensions.

Analyzing method comparison data with skew-normal measurement error models: incorporating generalized scale mixtures and varying degrees of freedom

Communications in statistics. Simulation and computation, Jun 24, 2024

Sri Lanka Anatomy Journal, Dec 30, 2022

Objective: The objective of this study was to analyse morphometric characteristics of the Peroneu... more Objective: The objective of this study was to analyse morphometric characteristics of the Peroneus Tertius (PT) and Extensor Digitorum Longus (EDL) muscles with important correlations and associations that may describe their functional significance. Material and Methods: The measurements PT and EDL muscles of cadavers were taken using a standard measuring tape. Minitab software Version 20.1 was used for statistical analysis methods of Spearman rho, Mood's Median test, Fishers Exact test, and Two-sample Ttest. Results: Of the 54 specimens of lower limbs, the mean length of origin of the PT from the fibula was 13.80±5.52 cm and 76% of the specimens with a separate origin of PT extended to its proximal half. The distal insertion of PT tendon was mostly as a single insertion into the base of the fifth metatarsal bone (55.55%). Spearman rho correlation value for the EDL muscle belly circumference with the number of intertendinous connections showed a mild negative correlation. Conclusions: Knowledge of morphometric characteristics and the variations of these muscles aid in clinical applications and suggest the use of radiological assessment of the muscle prior to surgical intervention.

Journal of Applied Mathematics, Statistics and Informatics, Dec 1, 2022

When implementing newly proposed methods on measurements taken from a human body in clinical tria... more When implementing newly proposed methods on measurements taken from a human body in clinical trials, the researchers carefully consider whether the measurements have the maximum accuracy. Further, they verified the validity of the new method before being implemented in society. Method comparison evaluates the agreement between two continuous variables to determine whether those measurements agree on enough to interchange the methods. Special consideration of our work is a variation of the measurements with the magnitude of the measurement. We propose a method to evaluate the agreement of two methods when those are heteroscedastic using Bayesian inference since this method offers a more accurate, flexible, clear, and direct inference model using all available information. A simulation study was carried out to verify the characteristics and accuracy of the proposed model using different settings with different sample sizes. A gold particle dataset was analyzed to examine the practical viewpoint of the proposed model. This study shows that the coverage probabilities of all parameters are greater than 0.95. Moreover, all parameters have relatively low error values, and the simulation study implies the proposed model deals with the higher heteroscedasticity data with higher accuracy than others. In each setting, the model performs best when the sample size is 500.

Medico-Legal Journal of Sri Lanka

Introduction: Sex determination using human skeletal remains is a challenging task for forensic p... more Introduction: Sex determination using human skeletal remains is a challenging task for forensic practitioners and foramen magnum is used at lesser extents for this purpose. The aim of this study was to determine the sex by evaluating the parameters of the foramen magnum in a Sri Lankan population using computed tomography (CT).Methods: CT images of 300 individuals aged between 20 to 60 years, comprising 146 males (49%) and 154 females (51%), obtained from the Radiology Department of National Hospital, Kandy, Sri Lanka were retrieved for the study. Four parameters of the foramen magnum, namely length, width, circumference, and area, were measured/calculated using the RadiAnt Dicom Viewer 2022.1 software and analyzed using SPSS 26 software.Results: The analyses indicated that all four measurements were significantly higher in males than in females. All the parameters showed positive correlations with each other. Discriminant function analysis indicated that length was the most dimorph...

Journal of Probability and Statistics, Dec 15, 2022

Method comparison studies mainly focus on determining if the two methods of measuring a continuou... more Method comparison studies mainly focus on determining if the two methods of measuring a continuous variable are agreeable enough to be used interchangeably. Typically, a standard mixed-efects model uses to model the method comparison data that assume normality for both random efects and errors. However, these assumptions are frequently violated in practice due to the skewness and heavy tails. In particular, the biases of the methods may vary with the extent of measurement. Tus, we propose a methodology for method comparison data to deal with these issues in the context of the measurement error model (MEM) that assumes a skew-t (ST) distribution for the true covariates and centered Student's t (cT) distribution for the errors with known error variances, named STcT-MEM. An expectation conditional maximization (ECM) algorithm is used to compute the maximum likelihood (ML) estimates. Te simulation study is performed to validate the proposed methodology. Tis methodology is illustrated by analyzing gold particle data and then compared with the standard measurement error model (SMEM). Te likelihood ratio (LR) test is used to identify the most appropriate model among the above models. In addition, the total deviation index (TDI) and concordance correlation coefcient (CCC) were used to check the agreement between the methods. Te fndings suggest that our proposed framework for analyzing unreplicated method comparison data with asymmetry and heavy tails works efectively for modest and large samples.

Prediction of Age Based on Development of Mandibular Third Molars in Sri Lankan Population

2022 6th SLAAI International Conference on Artificial Intelligence (SLAAI-ICAI)

Age estimation is fundamental to forensic expertise and clinical medicine. The third molar offers... more Age estimation is fundamental to forensic expertise and clinical medicine. The third molar offers one of the unique benefits that proceed over a more extended period. Demirjian's method is used to classify the third molar development based on eight stages. The stages were allocated a biologically weighted score for each gender. The main objective of this study is to predict the age of subadults based on the third molar development stages. Each third molar development stage was analyzed according to their side and gender. In this study,1643 left lower third molars and 1665 right lower third molars are considered for analysis, and the third molars' development stages were recorded in the age group from 10 to 28. Generalized Linear Mixed Model (GLMM), classification and regression tree algorithm (CART), Ridge regression, and Elastic net regression were used to predict the age. Results were validated using the cross-validation technique. Root mean squared error (RMSE), mean absolute error (MAE), and R-squared values were used to select the best model. There were significant differences between the male and female third molars, and there were no significant differences between the left and right lower third molars. Weighted Demirjian's stages and gender were the significant variables of the fitted models for predicting age. The best model for the prediction of age was the classification and regression tree algorithm (CART), which gave the highest accuracy (70.6%) with the minimum root mean squared error (RMSE = 2.27). Therefore, the classification and regression tree algorithm (CART) can be used to predict the age using the development stages of third molars.

International Journal of Forensic Odontology

The absence of dual career programmes in Sri Lanka cause premature retirement of student athletes... more The absence of dual career programmes in Sri Lanka cause premature retirement of student athletes from their elite athletic career or illustrious academic career. In most cases student athletes are pressurized to select an academic career which may promote a sedentary lifestyle and lead to a variety of health and social issues. However, implementation of dual career programmes requires a clear understanding of student-athlete motivation. The aim of the present study was to describe the motivation for dual careers of the undergraduate student athletes of two top ranked universities in Sri Lanka; University of Colombo and Peradeniya, using the Italian-Slovenian version of the Student-athletes' Motivation towards Sports and Academics Questionnaire(SAMSAQ-IS). SAMSAQ-IS is a self-administered questionnaire tested and validated in an environment more suitable to Sri Lanka consisting of 39 items. Student Athletic Motivation (SAM), motivation towards academic-related tasks (AM) and motivation to pursue a professional sport career (CAM) have been considered as three factors in this tool of assessment. Exploratory factor analysis, confirmatory factor analysis and Rasch analysis were applied to test the factor structure, reliability and validity of the SAMSAQ-IS. Two hundred and sixty six (266) student athletes participated in this study (males 63.5%). The mean ages were 22.98±1.39 years (females), and 22.94±1.7 years (males), participating in either team or individual sports. The majority were from the faculties of Science (28%), Management (18%), and Medicine (11%). All 39 items had a good reliability (Cronbach's alpha>0.7). The Rasch analysis showed that all infit and outfit statistics were within the range 0.5-1.5 and all items were productive for measurement. Gender disparity among 3 factors were noted where females showing more motivation for AM and male showing greater motivation for both Sam and CAM. However, overall positive motivation was observed in all factors. Overall SAM with CAM and SAM with AM had a significant positive correlation (p< 0.05). Overall, a high positive motivation was shown by student athletes despite the lack of an adequate support system. Gender differences in AM and CAM may be studied in depth when planning career transition programmes. Nevertheless this motivation may assist student athletes in overcoming some challenges in their progression and will positively impact the planning and implementation of career transition programmes, policies and guidelines at the national level.

The number of vehicles and road transportation increases rapidly daily. Hence the frequency of ro... more The number of vehicles and road transportation increases rapidly daily. Hence the frequency of road accidents and crashes also gradually increase with it. Analysing traffic accidents is one of the essential concerns in the world. Due to the considerable number of casualties and fatalities caused by those accidents, taking necessary actions to reduce road accidents is a vital public safety concern and challenge worldwide. Various statistical methods and techniques are used to address this issue. Hence, those statistical implementations are used for multiple applications, such as extracting cause and effect to predict realtime accidents. In this study, a United States (US) Countrywide car accidents data set consisting of about 1.5 million accident records with other relevant 45 measurements related to the US Countrywide Traffic Accidents were used. This work aims to develop classification models that predict the likelihood of an accident is severe. In addition, this study also consists of descriptive analysis to recognise the key features affecting the accident severity. Supervised machine learning methods such as Decision tree, K-nearest neighbour, and Random forest were used to create classification models. The predictive model results show that the Random Forest model performs with an accuracy of 83.95% for the train set and 80.69% for the test set, proving that the Random forest model performs better in accurately detecting the most relevant factors describing a road accident severity.

The heart is one of the essential organs in the human body. People are suffering from 'Myocardial... more The heart is one of the essential organs in the human body. People are suffering from 'Myocardial Infraction', 'Angioplasty' and 'Bypass surgery' or sudden death. Stress Echocardiography involves raising patients' heart rates through exercise. Then, take various measurements by pressuring the heart. Dobutamine can be used to pressure the heart, called Dobutamine Stress Echocardiography. Therefore, the main objective of this study is to propose models to predict cardiac diseases that can happen after giving the Dobutamine drug. This study was performed on a sample of 558 patients. This sample was taken by the Adult Cardiac Imaging and Hemodynamics Laboratories officers at the University of California, Los Angeles (UCLA). The study fits the statistical and machine learning models such as K-Nearest Neighbors (KNN), Naïve Bayes, Support Vector Machine (SVM), Decision Tree, Random Forest, Bagging methods with SVM, Gradient Boost, Extreme Gradient Boost (XG Boost), and Feedforward Neural Network (FFNN). Moreover, the hyperparametric tuning with the help of K-Fold Cross Validation techniques and Boosting methods were used to validate the fitted models and obtain better predictions. Furthermore, scaling methods such as Min-Max Scaling, Standard Scaling, and Quantité Scaling were used and handled the outliers to get better predictions without wasting much time. This study proposed five models corresponding to three diseases, sudden death, and any of these events. Myocardial infarction, angioplasty, bypass surgery, cardiac death, and any of these events can predict with 94.98%, 96.43%, 94.27%, 95.7%, and 84.44% accuracies.

Introduction: Sexual dimorphism is one of the most important implications in forensic investigati... more Introduction: Sexual dimorphism is one of the most important implications in forensic investigations and anthropological studies. Teeth are becoming a good source of material for gender determination. The canine is the most preferred tooth for gender determination because the canine is the strongest tooth in the oral cavity. Objectives: To investigate sexual dimorphism in permanent mandibular and maxillary canines of a Sri Lankan Sinhalese population, and to ascertain the most suitable dimension (labiolingual, mesiodistal and crown height) to determine the sex of an individual. Materials & Methods : The study was conducted using 384 dental casts (Males 192, Females 192) aged between 18 and 25 years in a sample of the Sri Lankan population. According to a selection criterion, casts were selected using a convenient random sampling technique. Mesio-distal, Bucco-lingual and Crown height of all the canines in the casts were measured using a digital vernier caliper accurate to 0.01 mm. Results : Statistical analysis was performed using Minitab 17 and SPSS (Version 21). Unpaired sample t-test, paired sample t-test and point-biserial correlation were used for data analysis. The present study revealed that males show larger mean dimensions of canine teeth than females. Out of all four canines, mandibular canines show highly consistent results for sexual dimorphism. Further, crown height is the best measurement to evaluate sexual dimorphism. Conclusion : It can be concluded that out of all the four canines, mandibular canines show highly consistent results for sexual dimorphism. Moreover, crown height is the best measurement to evaluate sexual dimorphism, in identifying an unknown individual.

Background: With the lack of specific treatment against COVID-19, Sri Lankans were seeking altern... more Background: With the lack of specific treatment against COVID-19, Sri Lankans were seeking alternative treatment options such as herbal medicines as preventive measures and treatment options against COVID-19. This study aimed to estimate the prevalence of such alternative treatment options usage by Sri Lankans during the pandemic and to assess the self-perceived effectiveness and adverse effects of herbal medicines from the participants' perception. Methods: An online cross-sectional survey was conducted among the general public. Data was collected using a questionnaire. A total of 804 participants were included in the study. Descriptive analysis was performed for all variables. A Chi-square test was performed to determine the association between the studied variables. Results: Among the participants, 90.4% reported using herbal medicines as preventive measures against COVID-19, and 86.7% used them to treat respiratory symptoms. Coriander and ginger were the most commonly used medicinal plants as preventives and in the treatment of respiratory symptoms. These herbs were perceived to be effective in alleviating respiratory symptoms by more than 85% of their users. A minority of the consumers (15.4%) experienced adverse effects associated with the use of herbal medicines as preventive measures. The use of herbal medicines as preventive measures was associated with the participant's age (p = 0.032) and education level (p <0.001). Conclusion: The study highlights the perceived effectiveness of some medicinal herbs in treating respiratory symptoms and recommends future research to isolate the compounds with potential pharmacological effects and conduct clinical trials to determine the effectiveness of the most commonly used plants.

Transformative Applied Research in Computing, Engineering, Science and Technology, 2025

J. Sci. Univ. Kelaniya, 2024

Statistics and Applications {ISSN 2454-7395), 2024

Analyzing method comparison data with skew-normal measurement error models: incorporating generalized scale mixtures and varying degrees of freedom

Communications in statistics. Simulation and computation, Jun 24, 2024

Sri Lanka Anatomy Journal, Dec 30, 2022

Journal of Applied Mathematics, Statistics and Informatics, Dec 1, 2022

Medico-Legal Journal of Sri Lanka

Journal of Probability and Statistics, Dec 15, 2022

Prediction of Age Based on Development of Mandibular Third Molars in Sri Lankan Population

2022 6th SLAAI International Conference on Artificial Intelligence (SLAAI-ICAI)

International Journal of Forensic Odontology

International Conference in Mathematics and Mathematics Education, 2019

Determining the gender identity from an unidentified human skeleton is of crucial importance to f... more Determining the gender identity from an unidentified human skeleton is of crucial importance to forensic anthropologists and bioarchaeologists. The femur is the largest bone in the human skeleton and therefore it is more likely to have forensic value. The main aim of this study is to develop a reliable and accurate method to determine the sex of contemporary Sri Lankan using measurements of the femur bone. A few classification methods are used to classify the gender of the human skeleton and those methods are compared in terms of their accuracy. A total of 22 measurements of head, neck, and other skeletal parts of both right bone and left bone of femur are used for this study. Discriminant analysis and classification methods are used to classify the gender of a femur. Only four measurements namely maximum length, subtrochanteric anteroposterior diameter, subtrochanteric medial-lateral diameter of the left bone, and mid shaft anteroposterior diameter of the right bone are significant in detecting the gender of the bone. In addition, the analysis reveals that all the characteristics of both the left side and right side of the femur bone have a similar variation. According to the mean value of all variables, male bone is larger than female femur bone. The accuracy of the sex identification of femur was obtained as 57.1%, 96.7%, 82.2%, and 88.4% from the discriminant analysis, logistic regression, K-Nearest Neighbor algorithm, and decision trees respectively. The classification methods performed well in determining the gender of femur bone than discriminant analysis. Hence, it can be concluded that the classification methods are more suitable to determine the gender of a femur of an unknown, mutilated, and dismembered human skeleton with higher accuracy.