Institutional-Repository, University of Moratuwa
Welcome to the University of Moratuwa Digital Repository, which houses postgraduate theses and dissertations, research articles presented at conferences by faculties and departments, university-published journal articles and research publications authored by academic staff. This online repository stores, preserves and distributes the University's scholarly work. This service allows University members to share their research with a larger audience.
 Research Publications |  Thesis & Dissertation |  E- Books |
 UoM Journal Publications |  UoM Conference Proceedings |  Articles published in Scimago's Q1 journals |  UoM Research Reports |  Other Articles authored by UoM staff |
Recent Submissions
item: Conference-Full-text
Comparative analysis of resampling techniques on class imbalance in body shaming phrase detection
(IEEE, 2024) Puvanendran, R; Wijikumar, P; Rathnayaka, T; Thilakarathna, K; Jayasiri, P; Roopasinghe, H; Kavishan, M; Jeyamohan, M; Thurshikan, K
In this present era, body shaming has pervasive and detrimental effects on individuals’ psychological and physical well-being. Hence, a profound significance in its potential to elucidate and address the textual based body shaming phrases is needed. Natural Language Processing (NLP) and Machine learning are employed to detect body shaming phrases. This research paper presents the machine learning methodology for the detection and classification of body shaming phrases. The standard TF-IDF has been employed for feature extraction. The SVM classifier exhibited an exceptional accuracy of 98.5%. In order to improve performance and address the issue of overfitting, the dataset is subjected to resampling procedures, which are carefully applied to ensure the appropriate selection of the model. The methodology encompasses four resampling strategies for achieving data balance, namely SMOTE, ADASYN,
SMOTE-Tomek and SMOTE-ENN. The efficacy of the suggested methodology is evaluated by comparing it against five machine learning classifiers, The findings suggest that SVM exhibit superior performance compared to other models in both the SMOTE and SMOTE-Tomek balanced datasets. Specifically, SVM achieves accuracy, recall, precision, and F1-score scores of
99.1%, 99.59%, 99.18%, and 99 .38% respectively. The study concludes that the utilization of resampling approaches, specifically SMOTE and SMOTE-Tomek enhances the model’s performance.
item: Conference-Full-text
The Impact of construction industry supply chain risk management for imported items due to the economic crisis; contractors’ perspective in Sri Lanka
(IEEE, 2024) Thejani, DN; Vithana, NDI
Supply chain management integrates suppliers, manufacturers, retailers, and warehouses to efficiently produce, distribute, and maintain products at the right times and quantities. The Sri Lankan construction industry faces unexpected risks due to the economic crisis, particularly in the supply chain for imported items. This study investigates these risks, and their impact on contractors, and aims to identify effective risk management strategies to prevent such disruptions. While achieving the objectives, both quantitative and qualitative approaches were followed. The data for the study was collected through closed-ended questions using a survey and semi-structured interviews were carried out to find further information. The quantitative analysis has been done using the Relative Importance Index (RII) and qualitative analysis has been done using content analysis. Based on this study, identified that the economic crisis of Sri Lanka has impacted the supply chain for imported items in the construction industry such as mainly the impact of local taxes and charges, price increasing rapidly for import items, and due to that it has affected the constructors such as cost overrun, loss of profit. The study suggests that Sri Lanka should focus on manufacturing quality and standard items to mitigate the impact of the current economic crisis and prevent future similar situations.
item: Conference-Full-text
SimuSIL: a software-in-the-loop simulator for cooperative driving in intelligent transportation systems
(IEEE, 2024) Kulathunga, R; De Silva,O; Abeywickrama, D; Vithanage, T; Dias, D
We present SimuSIL, a software-in-the-loop (SIL) simulation platform using open-source tools to capture the behavior of connected vehicles in intelligent transportation system (ITS) applications. Connected cooperative vehicles use vehicleto-everything (V2X) communication to enable ITS. Real-world testing of ITS applications involves significant risks. Simulating true vehicle dynamics and real-time behavior is challenging in discrete time step environments. We extend the Veins platform, integrating SUMO and OMNet++, with Webots, a simulator for robots and vehicles. Webots implements engine control unit (ECU) models in software to mimic physical vehicle behavior. We validate SimuSIL with a car-following application using Webots and demonstrate its ability to capture communication anomalies like packet losses. Additionally, we demonstrate fuel consumption and CO2 emissions in autonomous eco-driving versus human driving.
item: Conference-Full-text
Evaluation of satellite rainfall estimates for surface runoff modelling in the Maha Oya Basin, Sri Lanka
(IEEE, 2024) Suraweera, B; Gunawardhana, L; De Silva, K; Rajapakse, L
Temporal and spatial variations in rainfall are inherent in tropical monsoonal catchments with varying topographical characteristics, causing corresponding variations in river runoff. Global satellite rainfall estimates offer advantages in mitigating scale discrepancies and overcoming the critical requirement of reliable rainfall data sources in regions with limited ground-based observations for hydrological studies. This study comprehensively analyses the applicability of high-resolution satellite rainfall estimates (SREs) for surface runoff modeling. The efficacy of satellite estimates necessitates rigorous evaluation due to intrinsic biases, requiring bias correction. Hydrological simulations conducted using bias-corrected SREs reveal significant differences in model performance, emphasizing the importance of accurate rainfall inputs for reliable streamflow predictions. Overall, this research addresses the research gap in utilizing satellite rainfall estimates for hydrological modeling, particularly in regions with diverse terrain conditions and tropical monsoonal climates. CHIRPS, AgERA5, and PERSIANN rainfall estimates were selected considering their applicability in recent research studies. CHIRPS and AgERA5 rainfall estimates follow the monsoonal climate signal within the Maha Oya basin, fitting to ground-based rainfall observations. Further, CHIRPS estimates generate the best runoff simulations compared to other SREs, resulting in coefficient of determination of 0.53 and 0.57 for model calibration and validation.
item: Conference-Full-text
Mitigation of distributed harmonics in LV/MV grid using grid-tie inverters associated with Solar PV
(IEEE, 2024) Navaratne, MAUS; Maneesha, KDJ; Madhuthisari, HAP; Munasinghe, VCG
The presence of harmonics in power systems causes issues such as reducing component life, inaccurate energy measurements, overheating components, etc. The harmonics are typically handled at the load end and the consumers, typically the large scale industries are responsible for mitigating the harmonics produced. But With the increasing use of power electronic devices by residential customers, increasing penetration of gridconnected Distributed Energy Resources such as solar PV and increasing small-scale industries connected to the low voltage network increases the presence of harmonics in the medium voltage (MV) and low voltage (LV) network. The utility can not expect individual residential customers to mitigate these harmonics due to the high cost associated with the filtering techniques. Installing filters by the Utility also requires high capital costs. The distributed nature of the harmonic injection in MV/LV network also increases the complexity in introducing the filters into the system. This paper proposes a solution to mitigate distributed harmonics in LV/MV systems using the Grid Tie Inverters already available with distributed solar PV systems.