LNLakshika S. Nawarathna

Lakshika S. Nawarathna

*This is an algorithmically generated author record based on publicly available information.

Current Organizations / Institutes

The organizations/institutes from the scholar's latest article

Previous Organizations / Institutes

The organizations / institutes from the scholar's other articles in the past ten years


Lakshika S. Nawarathna has a total of 96 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • Open Access

    Knowledge, attitudes, and practices towards COVID-19 and their sociodemographic associations among communities in selected districts of Sri Lanka

    by H. S. Jayaneththi, L. S. Nawarathna, M. E. Eremeeva, R. S. Rajakaruna

    Coronavirus disease 2019 (COVID-19), caused by the SARS-CoV-2 virus, created a profound global health crisis that reshaped health systems worldwide. Compliance with preventive measures such as mask-wearing, hand hygiene, and vaccination is influenced by people’s knowledge, attitudes, and practices (KAP) in Sri Lanka. Understanding these dimensions and the factors associated with them is vital for effective pandemic preparedness and control. This study examined the KAP of Sri Lankans towards COVID-19, explored sociodemographic predictors, and compared the performance of Binary Logistic Regression (BLR) and Random Forest (RF) algorithms in classifying KAP levels. Responses were collected from 80 adults across four districts using an online questionnaire adapted from validated tools comprising 48 items. Confirmatory Factor Analysis (CFA) was employed to preliminarily assess and refine the measurement structure, while BLR and RF were applied to assess predictors. Results showed that 67.5% of respondents demonstrated good knowledge, 63.8% good practices, and 70.0% a positive attitude towards COVID-19. Education was significantly associated with all three KAP dimensions in the univariate analyses, while age and education were associated with knowledge and attitude, and education with practice. In the multivariable BLR models, Advanced Level (A/L) education was significantly associated with higher odds of good knowledge and attitude, while no predictor was statistically significant for practice. CFA supported a reduced measurement structure in which a subset of items showed satisfactory factor loadings across the three KAP dimensions. A comparison of classification performance revealed that BLR demonstrated higher Area Under the Curve (AUC), accuracy, recall, and F1-scores than RF across all three KAP dimensions, although RF achieved higher precision for knowledge and practice. These findings underscore the critical role of education in shaping public responses and suggest that different KAP dimensions may benefit from different analytical approaches. The study offers valuable insights for policymakers and health authorities in designing targeted health education campaigns to address sociodemographic disparities and strengthen community resilience. Despite limitations of sample size and convenience sampling, the research provides a foundation for future large-scale, representative, and longitudinal studies on KAP and similar contexts.

  • Open Access

    Evaluating the Efficacy of Machine Learning Algorithms for Customer Churn Prediction in The Insurance Industry

    by K.M.S.S.B. Kulasekara, L.S. Nawarathna

    In the insurance industry, large-scale data from customer transactions is generated daily, making it more cost-effective to retain existing clients than to acquire new ones due to customer churn. This study aims to identify the best classification algorithm for predicting customer churn, determine the characteristics of churning customers, and model the time until churn. After data preprocessing and addressing dataset imbalance using the Synthetic Minority Oversampling Technique (SMOTE), various predictive models, including logistic regression, naive Bayes, decision trees, random forests, neural networks, extra trees, boosting algorithms, and heterogeneous ensemble learning methods, were applied. K-fold cross-validation was utilized for hyperparameter tuning and preventing overfitting. The results, evaluated using the confusion matrix and Area Under Curve (AUC), showed that Extreme Gradient Boosting (XGBoost) and Light Gradient Boosting Machine (LGBM) classifiers achieved the highest accuracy at 88.57% and 88.53%, respectively, with precision, recall, and F-score also used for assessment. Survival analysis revealed that approximately 80% of policyholders survived the first 200 months, with “having children” being the most significant predictor of customer churn. Unlike standard classification, survival analysis enables modeling of both churn likelihood and expected time-to-churn, accounting for censored observations. This provides actionable insights for intervention timing, a critical advantage in customer retention planning

  • Urban Temperature Dynamics in Sri Lanka: A Comparative Machine Learning and Time-series Analysis Across Colombo, Kurunegala, and Puttalam Districts

    by R. A. K. P. Rupasinghe, Lakshika S. Nawarathna

    Rapid urbanisation and land‑use change have intensified temperature variability in tropical cities, creating new challenges for climate resilience. Sri Lanka, which experiences heat stress across all districts, including Colombo (wet zone), Kurunegala (intermediate zone), and Puttalam (dry zone), requires reliable daily temperature forecasts to support agriculture, energy management, and public health planning. Using 4,916 daily meteorological observations (2010–2023), this study compares the performance of ARIMAF time‑series models with Multiple Linear Regression, Support Vector Regression (SVR), Random Forest, and Extreme Gradient Boosting (XGBoost). Among these approaches, SVR with the radial basis function kernel achieved the highest predictive accuracy (R² = 0.66 – 0.80; MAE = 0.34–0.44°C; RMSE = 0.45–0.57°C), while Random Forest and XGBoost achieved R² values between 0.60–0.75, outperforming ARIMAF (R² = 0.45–0.55), which showed weaker predictive capacity under nonlinear conditions. Feature importance analysis confirmed apparent temperature and reference evapotranspiration as dominant predictors across all districts, contributing >30% of the variance explained by the models. Unlike earlier Sri Lankan studies that emphasised seasonal or monthly analyses, this work provides the first systematic comparison of multiple machine learning and time‑series approaches for daily forecasting using high‑resolution daily data across all three districts. Although the limited availability of external datasets constrained external validation, methodological consistency was ensured through chronological train–test splitting and standardised evaluation metrics. The findings establish a robust framework for tropical urban temperature forecasting, offering practical insights for climate‑resilient urban planning and policy development.

  • Open Access

    Evaluating Agreement Among Multiple Methods With Replicated Measurements Using Scale Mixtures of Skew‐Normal Measurement Error Models

    by Jeevana Duwarahan, Lakshika S. Nawarathna

    This paper introduces a robust parametric approach for assessing the agreement among multiple measurement methods when dealing with replicated data from a continuous variable. Many existing methodologies in the literature rely heavily on the normality assumption and often employ mixed‐effects models. However, in practical scenarios, these assumptions may not hold true due to the presence of skewness and heavy‐tailed distributions. Additionally, inherent measurement errors can further complicate the use of mixed‐effects models. To overcome these issues, we propose a multivariate measurement error model that assumes scale mixtures of skew‐normal distributions for the unobserved true covariates and scale mixtures of normal distributions for the errors. This formulation flexibly accommodates both skewness and heavy‐tailed characteristics within the data, allowing for different degrees of freedom for the true covariate and error distributions. An expectation‐conditional‐maximization algorithm is developed to estimate the parameters of this novel model. Furthermore, we employ the concept of the probability of agreement to evaluate agreement among the various pairs of measurement methods. To validate the efficacy of our proposed model, we conducted extensive simulation studies involving various sample sizes (50, 100, and 200) and varying skewness parameters (1, 5, and 10). Moreover, we demonstrate the practical applicability of our methodology by applying it to real tumor data. Our findings conclusively demonstrate the effectiveness of the proposed model in the analysis of replicated method comparison data across multiple methods, particularly in situations involving measurement errors, skewness, and heavy‐tailed distributions.

  • Naturally occurring iron-aluminum oxides and oxy-hydroxides coated sand for removing cationic dyes from contaminated water

    by N. H. Koralegedara, D. Gamage, I. Rathnayake, K. Jayasoma, C. Chandraratne, R. Chandrajith, L. S. Nawarathna

  • Open Access

    A model for predicting confirmed leptospirosis cases in Sri Lanka

    by W. E. M. D. T. Ekanayake, L. S. Nawarathna

    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 (non-parametric). 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). We found that the ARIMA model was the best-fitting model for Galle, Mullaitivu, Batticaloa, and Puttalam. The exponential regression model was the best-fitting model for Kalmunai, while the polynomial regression model was the best-fitting model for Matale and Jaffna. The generalised additive model was the best-fitting model for Gampaha, Hambantota, Matara, Mannar, Vavuniya, Trincomalee, Anuradhapura, Badulla, Monaragala, and Ratnapura. In contrast, the median-based linear model was the best-fitting model for Colombo, Kalutara, Kandy, Nuwara Eliya, Kilinochchi, Ampara, Kurunegala, Polonnaruwa, and Kegalle. The fitted models were then used to forecast leptospirosis incidents in Sri Lanka from 2024 to 2026. For future research, we recommend Bayesian hierarchical models to identify high-risk clusters of leptospirosis in Sri Lanka.


  • Exploring the impact of e-learning instructional designs on students’ learning approaches and knowledge gain

    by Chantha K Jayawardena, Ruwinee T Dissanayake Mudiyanselage, Kosala Marambe, Lakshika S Nawarathna

    Designing educational materials for digital resources that accommodate different learning styles and educational backgrounds is challenging due to the complexity of student engagement with content. This study examined the impact of E-learning instructional designs on students’ knowledge gain, focusing on learning approaches and perceived usefulness. The 'Virtual Histology Lab.’ (VHL), a self-learning E-resource, was developed to deliver theoretical knowledge on microscope slide preparation, incorporating cognitive scaffolding strategies such as contextualized trigger questions, map-based visualizations, and multimedia elements grounded in sound pedagogical principles. Fifty-one students with no prior exposure to slide preparation participated in the study. Knowledge gain was assessed through pre-test and post-test scenarios, while three questionnaires evaluated learning approaches, perceived usefulness, ease of use, and student perspectives on the VHL’s content. The findings revealed that 78% of students predominantly adopted a deep learning approach, while a smaller proportion exhibited a surface approach. No significant differences were observed between the pre-test and post-test scores of students with deep or surface learning approaches. However, all students demonstrated a statistically significant increase in post-test scores, indicating substantial knowledge gains across diverse learning styles. This study shares instructional design strategies to accommodate different learning styles, increase knowledge gain, and promote meaningful learning experiences.


  • Open Access

    Assessing groundwater quality in the northwestern coastal area of Sri Lanka using multivariate statistical techniques

    by N. M. A. N. Weerasinghe, L. S. Nawarathna, H. M. T. G. A. Pitawala

    The objective of the present study was to assess groundwater quality in the northwestern coastal region of Sri Lanka by analyzing hydrogeochemical data using multivariate techniques. The study examined the relationship between water quality and sources influencing groundwater quality and proposed useful statistical tools for water resource management. Descriptive, comparative, correlation, factor analysis (FA), and discriminant analysis (DA) were among the techniques used. The water quality in terms of electrical conductivity (EC), total dissolved solids (TDS), and dissolved nitrate and phosphate concentrations of the northwestern coastal area, where different aquifer conditions prevailed, were compared, and differences in the mean concentrations of the above parameters were identified. The study also employed ridge regression and support vector machines (SVM) to predict nitrate levels and validate the results using a 70:30 train/test split. The analysis found that the mean concentrations of EC, TDS, and HCO3- are higher in the sandy aquifer area, while Fe2+ and PO43- are higher in the regolith aquifer area. EC, TDS, PO43-, and HCO3- in the hard-rock aquifer area exceed permissible limits. Factor analysis identified four factors contributing to pollution, and a linear discriminant model classified all observations with 99.9% accuracy. The water quality index for Kalpitiya, Vanathavillu, and Katana-Negombo areas was 68.1, 99.2, and 287.9, respectively, while the World Health Organization (WHO) cutoff was 50. The use of multivariate statistical techniques demonstrated the value of analyzing hydrogeochemical data to gain valuable insights into the complex interplay of factors affecting groundwater quality. The SVM model, using a linear kernel, was identified as the best model to predict nitrate levels. A combination of statistical analyses reveals that, with the exclusion of a few wells, the groundwater in the study area is polluted and inappropriate for human consumption or culinary purposes. Nonetheless, the inhabitants persist in utilizing it as their primary source of drinking water. This study reveals the combination of statistical analyses is useful for the management of groundwater resources.


  • Open Access

    Multivariate statistical techniques for the evaluation of hydrogeochemical data from areas with chronic kidney disease of unknown etiology (CKDu) in Sri Lanka

    by Rathnayake Hmad, Lakshika S. Nawarathna, Rohana Chandrajith

    Chronic Kidney Disease of unknown etiology (CKDu) ranked as one of the most attended public health issues in Sri Lanka. The etiology of the disease is believed to be environmentally induced as it appears in geographically discreet regions of the dry zone of Sri Lanka. Girandurukotte and Wilgamuwa are two regions where CKDu is more prevalent. This study performs a comparative analysis to separately identify the quantitative differences of hydrogeochemical parameters of groundwater samples consumed by the patients and non-patients in two regions. Besides, groundwater quality data of 421 wells from Girandurukotte and Wilgamuwa were used for the analysis. Out of 421 samples, 310 samples were collected from wells where patients were reported. Descriptive statistics, Correlation analysis and Factor analysis were performed. Results of comparative analysis conclude that means of iron (Fe) content in Wilgamuwa CKDu (371 μg/L) and non-CKDu (731μg/L) exceeds the WHO permissible limits. The mean ion content of F-, Cl-, NO 3-, SO 4 2-and Na + in the water wells consumed by the CKDu patients in the Wilgamuwa region is higher in concentration than non-CKDu wells. Out of major elements, Mg 2+ , Ca 2+ , K + , Na + , NO 3-and hardness contents are higher in Wilgamuwa CKDu wells compared to the Girandurukotte CKDu wells. Trace elements such as As, Cd, and Pb do not exceed the WHO recommended limits in both regions. Generally, except for Fe ion concentration, almost all the other ion concentrations are within the WHO recommended limits. Correlation analysis reveals that Total Dissolved Solids (TDS) is highly positively correlated with alkaline earth metals Mg 2+ (0.93), Ca 2+ (0.86) and Sr 2+ (0.85) in Wilgamuwa CKDu water sources. Also, Electrical Conductivity (EC) is highly positively correlated with Ca 2+ (0.84), Ni (0.82) and Sr (0.81) in Wilgamuwa non-CKDu water sources. Factor analysis was applied to identify the chemical combinations in each cluster. Ni, Ca 2+ , Sr 2+ and Mg 2+ appear as a factor in Wilgamuwa non-CKDu water sources. Hence, alkaline earth metals are chemically associated and appear as a factor combination in Wilgamuwa non-CKDu water sources. Also, Alkalinity, Hardness, Ca 2+ and Sr 2+ were found as a factor in Ginnoruwa CKDu samples.