Browsing by Author "Banujan, Kuhaneswaran"

Now showing 1 - 4 of 4

An Ensemble Machine Learning Approach for Stroke Prediction
(Department of Industrial Management, Faculty of Science, University of Kelaniya Sri Lanka, 2022) Premisha, P.; Prasanth, Senthan; Kanagarathnam, Mauran; Banujan, Kuhaneswaran
Nowadays, one out of four people above 25 will suffer from a stroke. Especially this year, with the highest count of around 13.7 million people discovered with stroke for the first time. Out of 13.7 million, 5.5 million were fatalities. This was stated in a recent WHO study. It is estimated that if no action is taken, the number of fatalities will rise to 6.7 million yearly. The pandemic situation of COVID-19 will play a significant cause in the expanded death rate of stroke. Even for adults and patients with minor risk factors affected by stroke rather than in previous years. This study predicts the impact level of stroke with the development of an ensemble model by combining the various classifiers performed well in isolation. Predicting the stroke status in patients would help the physicians determine the prognosis and assist them in providing the targeted therapy in a limited time. During this study, an ensemble model was built by considering the base, bagging, and boosting classifiers: Support Vector Machine, Naïve Bayes, Decision Tree, Logistic Regression, Artificial Neural Network, Random Forest, XGBoost, LightGBM, and CatBoost. The dataset consists of 5110 patient details, along with 12 attributes that were analyzed in this research. The final ensemble model was developed by carrying out the methodology in two phases. During the first and second phases, the classifiers mentioned above were trained without hyper-parameter tuning and with hyperparameter tuning and tested against the fundamental evaluation matrices. During each phase, the classifier that produces the highest classification accuracy is discovered from the base, bagging, and boosting categories. From the results obtained, the final ensemble model was constructed using the Max Voting approach, which yielded an accuracy of 95.76%.
A Machine Learning Influenced Recommendation System for Predicting the Rainfall and Price for Crops in Badulla District
(Faculty of Computing and Technology, University of Kelaniya Sri Lanka, 2022) Nandasiri, K.P. Sasindu Madushan; Banujan, Kuhaneswaran; Kumara, B.T.G.S.; Jayasinghe, Sadeeka; Ekanayake, E.M.U.W.J.B.; Senthan, Prasanth
Every day, agriculture becomes more vital to the global economy. Daily population expansion necessitates substantial crop output for human existence. But as the population has increased, human activity has also altered the environment. Therefore, it has resulted in challenges with weather forecasting, which is crucial for crop planting in the agricultural sector. Thus, the globe needs a method to forecast agrarian weather. In addition, it is highly advantageous for farmers to understand the production rate they can achieve and the price range they may expect for their efforts. As a result, Machine learning technologies have become unique and fashionable in the agricultural industry due to their ability to provide accurate farming predictions. Selecting suitable plants for planting has evolved into a necessity. This study focuses on the application of machine learning to estimate the optimal crop for a given period. In this work, the author addresses the beginning part of the study: precipitation prediction under the weather forecast and pricing forecast. The authors have employed six distinct machine-learning models to forecast rainfall and crop prices.
The Role of Social Media (Twitter) in Analysing Home Violence: A Machine Learning Approach
(Department of Industrial Management, Faculty of Science, University of Kelaniya Sri Lanka, 2023) Adeeba, Saleem; Banujan, Kuhaneswaran; Kumara, B.T.G.S.
Home Violence (HV) has been a persistent issue across the globe, transcending economic status and cultural boundaries. The COVID-19 pandemic has further exacerbated this problem, bringing it to the forefront of public discourse. This study aims to analyse the impact of HV by utilising Twitter data and Machine Learning (ML) techniques, categorising tweets into three groups: (i) HV Incident Tweets, (ii) HV Awareness Tweets, and (iii) HV Shelter Tweets. This categorisation provides several advantages, such as uncovering new or hidden evidence, filling information gaps, and identifying potential suspects. Over 40,000 tweets were collected using the Twitter API between April 2019 and July 2021. Data pre-processing and word embedding were performed to prepare the data for analysis. Initially, tweets were categorised into HV Positive (containing relevant information) and HV Negative (noise or unrelated content) groups. Manually labelled tweets were used for training and testing purposes. Machine learning models, including Support Vector Machines (SVM), Naïve Bayes (NB), Logistic Regression, Decision Tree Classifier, Artificial Neural Networks (ANN), and BERT+LSTM, were employed for this task. Subsequently, HV Positive tweets were classified into the three aforementioned categories. Manually labelled tweets were again used for training and testing. Models such as Tf-IDF+SVM, Tf- IDF+Decision Tree, Tf-IDF+NB, and GloVe+LSTM were utilised. Several evaluation metrics were used to assess the performance of the models. The study’s results provide important new understandings of the prevalence, patterns, and causes of HV as they are reported on social media and how the general population reacts to these problems. The research clarifies how social media may help spread knowledge, provide assistance, and link victims to resources. These insights can be instrumental in informing policymakers, non-profit organisations, and researchers as they work to develop targeted interventions and strategies to address HV during and beyond the COVID-19 pandemic.
Stack Ensemble Model to Detect the Stress in Humans by Considering the Sleeping Habits
(Faculty of Computing and Technology, University of Kelaniya Sri Lanka, 2022) Kanagarathnam, Mauran; Premisha, P.; Prasanth, Senthan; Banujan, Kuhaneswaran; Kumara, B.T.G.S.
Recently, one of the big challenges encountered by humans is experiencing and managing stress. Beyond the age restriction, people of all ages, from teenagers to seniors, experience issues as a result of stress. Acute and chronic stress are the two main categories of stress. Acute stress is a typical human response that aids in your body’s adaptation to a new situation. In actuality, this form of stress has positive effects. However, the second type of stress, chronic stress, is a crucial type of stress, and this study focused on determining the stress level of this type in advance. This research examined eight attributes related to chronic stress to investigate the chosen person’s sleeping patterns. The Kaggle website provided the dataset that was used in this study. The user’s snoring range, body temperature, limb movement rate, blood oxygen levels, eye movement, number of hours of sleep, heart rate, and stress levels (0-low/normal, 1-medium low, 2-medium, 3-medium high, 4 - high) were all taken into account. The stack ensemble approach was utilized with two levels during this approach. At level 0, the classifiers such as Random Forest, Decision tree, K-nearest neighbour, and XGBoost were considered. At level 1, as a Metamodel, Logistic regression was adopted. Moreover, the predictions obtained from the level 0 models added an additional attribute to the original dataset and fed it to the level 1 model as a new training dataset. Additionally, five folds of fold cross-validation were performed along with the basic assessment to validate further the model for various ratios of training and testing data. Following the cross-validation, the model’s mean accuracy obtained for RF, DT, KNN, XGB and stack ensemble models. From the results discovered, it was represented that the combined model (stack ensemble model) produced more precise results rather than the models considered in isolation.