Browsing by Author "Rajapaksha, R."

Now showing 1 - 3 of 3

Non-invasive diagnostic approach for diabetes using pulse wave analysis and deep learning
(MDPI, 2024) Gunathilaka, H.; Rajapaksha, R.; Kumarika, T.; Perera, D.; Herath, U.; Jayathilaka, C.; Liyanage, J.; Kalingamudali, S.
The surging prevalence of diabetes globally necessitates advancements in non-invasive diagnostics, particularly for the early detection of cardiovascular anomalies associated with the condition. This study explores the efficacy of Pulse Wave Analysis (PWA) for distinguishing diabetic from non-diabetic individuals through morphological examination of pressure pulse waveforms. The research unfolds in four phases: data accrual, preprocessing, Convolutional Neural Network (CNN) model construction, and performance evaluation. Data were procured using a multipara patient monitor, resulting in 2000 pulse waves equally divided between healthy individuals and those with diabetes. These were used to train, validate, and test three distinct CNN architectures: the conventional CNN, Visual Geometry Group (VGG16), and Residual Networks (ResNet18). The accuracy, precision, recall, and F1 score gauged each model’s proficiency. The CNN demonstrated a training accuracy of 82.09% and a testing accuracy of 80.6%. The VGG16, with its deeper structure, surpassed the baseline with training and testing accuracies of 90.2% and 86.57%, respectively. ResNet18 excelled, achieving a training accuracy of 92.50% and a testing accuracy of 92.00%, indicating its robustness in pattern recognition within pulse wave data. Deploying deep learning for diabetes screening marks progress, suggesting clinical use and future studies on bigger datasets for refinement.
Pilot study for non-invasive diabetes detection through classification of photoplethysmography signals using convolutional neural networks
(University of Kelaniya, 2024) Gunathilaka, H.J.; Rajapaksha, R.; Kumarika, T.; Perera, D.; Herath, U.; Jayathilaka, C.; Liyanage, J.A.; Kalingamudali, S.R.D.
Diabetes is a chronic disorder affecting vascular health, often altering pulse wave characteristics. Traditional pulse wave analysis (PWA) methods face challenges such as variability and complexity of signals. This study aims to overcome these limitations by leveraging deep learning models for more accurate and efficient classification. The methodology used in this study involves four key steps: data collection, data preprocessing, Convolutional Neural Network (CNN) model development, and model evaluation. Primary data were collected using a multipara patient monitor, including finger photoplethysmography (PPG) signals, blood pressure, mean arterial pressure, oxygen saturation, and pulse rate. Single pulse wave cycles from 60 healthy individuals and 60 patients with type 2 diabetes underwent preprocessing. The CNN model was trained using 50 PPG images from each group and achieved a training accuracy of 92%. The prediction capability of the model was evaluated using 20 unseen images, comprising 10 healthy and 10 diabetes PPG images. It attained a 90% overall test accuracy in distinguishing between PPG images of individuals with diabetes and those who are healthy. These findings suggest that CNNbased analysis of PPG signals provides a precise, non-invasive tool for diabetes screening. To further enhance accuracy, future studies should focus on increasing the dataset size and performing hyperparameter tuning to optimize the CNN model.
Temporal cross-validation in forecasting: A case study of COVID-19 incidence using wastewater data
(Quality and Reliability Engineering International, 2024-11) Lai, M.; Wulff, S. S.; Cao, Y.; Robinson, T. J.; Rajapaksha, R.
Two predominant methodologies in forecasting temporal processes include traditional time series models and machine learning methods. This paper investigates the impact of time series cross-validation (TSCV) on both approaches in the context of a case study predicting the incidence of COVID-19 based on wastewater data. The TSCV framework outlined in the paper begins by engineering interpretable features hypothesized as potential predictors of COVID-19 incidence. Feature selection and hyperparameter tuning are then utilized with TSCV to identify the best features and hyperparameters for optimal model performance given a specific forecast horizon. While evidence supporting the utility of TSCV for auto-regressive integrated moving average model with exogenous variables (TS-ARIMAX) forecasts is lacking in this study, such an approach proves advantageous for gradient boosting machine forecasts (TS-GBM). In Wyoming, for instance, TS-GBM had a 34.9% improvement compared to naïve predictions, whereas GBM without TSCV only had a 15.6% improvement. However, TSCV also enhances interpretability for both TS-ARIMAX and TS-GBM models as this approach selects specific features, such as lagged values of COVID-19 cases, based on forecast performance and forecast length. Future research should work to explore the influence of stationarity and model averaging on the performance of TSCV in forecasting applications.