3830 research outputs found
Sort by
A Comparative Study on Utilizing Machine Learning and Ensemble Learning to Classify to Predict Air Quality
No full textIn this study, I will address air quality forecasting by employing machine learning algorithms to estimate the hourly concentrations of air pollutants (Such as, ozone, particle matter (PM2.5), and sulphur dioxide). Machine learning, one of the most common approaches, is capable of effectively training a model on massive amounts of data by employing large-scale optimization algorithms. Although there are several works that use machine learning to forecast air quality, earlier researches are limited to many years of data and simply train conventional regression models (linear or nonlinear) to predict hourly air pollution concentrations. In this paper, we offer updated models for predicting hourly air pollution concentrations based on previous days' meteorological data by structuring the prediction across 24 hours as a multi-task learning (MTL) issue. This allows us to choose a decent model using various regularization strategies. Air pollution has been a big issue for the general population and governments all around the world
Exploring Graph-Based Machine Learning Techniques for Transaction Fraud Detection: A Comparative Analysis of Performance
No full textThis report investigates graph-based fraud detection methods, focusing on credit card transactions to tackle the increasing complexity of financial fraud. It explores various machine learning models viz. Isolation Forest, Random Forest, Graph Autoencoders (GAE), and Graph Convolutional Networks (GCN), assessing them based on evaluation metrics. The Random Forest emerges as a robust model with consistent high performance, while the Isolation Forest shows minimal effectiveness. The GAE and GCN demonstrate potential, especially with hyperparameter tuning. Significant improvements in accuracies were observed post-tuning, particularly with the GCN model, showcasing the importance of model optimization. The research acknowledges the challenges of acquiring graph structured data, real-time analysis, adaptive fraudsters, and data privacy in implementing graph-based fraud detection. Conclusively, the study endorses graph-based methods as a formidable approach to enhance fraud detection, emphasizing continuous research and development to address existing challenges and improve system scalability, efficiency, and security. Accuracies obtained posttuning are notably high for Random Forest and GCN, indicating their effectiveness in fraud detection scenarios
E-commerce Sales Forecasting Using Machine Learning Algorithm
No full textBusinesses looking to maximise inventory, marketing tactics, and overall operational efficiency must consider e-commerce sales forecasts when making strategic decisions. The goal of this project is to determine the best method for predicting e-commerce sales by developing, assessing, and contrasting three time series machine learning models: ARIMA, Facebook (FB) Prophet and LSTM. The goals of the study are to prepare the dataset, create the model, tune the hyperparameters, and evaluate the performance. It explains how complicated e-commerce sales trends are and highlights the need for improved models or tactics to identify subtle patterns. Out of all the 3 models Facebook (FB) Prophet outperformed ARIMA and LSTM with decent evaluation matrix score like (RMSE): 273.79, (MSE): 74965.26, (MAE): 221.24. It also identifying intricate patterns and offering insightful analysis of e-commerce sales data. However the FB prophet also fails to give accurate e-commerce sales predictions like the ARIMA and LSTM models
Assessing Commercial Wearables in Predicting Physical Activity: A Case Study of Apple and Fitbit
No full textThis study explored the predictive accuracy of Apple Watch and Fitbit in tracking physical activity, employing advanced machine learning models, feature engineering, ensemble learning, and interpretable machine learning. Using Microsoft Power BI, an interactive dashboard was also constructed to analyze user demographics, physiological metrics, and activity patterns. The machine learning models, including Logistic Regression, Naïve Bayes, Decision Tree, Random Forest, LightGBM, XGBoost, CatBoost, and Artificial Neural Network (ANN) were comprehensively evaluated using multiple metrics. Interpretability was enhanced through Shapley values, unravelling the contribution of features to classification results. Stacking models reveal insights into their performance compared to individual models. The result generally showed that the single LightGBM model was better compared to other models and stacking. Furthermore, the dashboard insights provide a detailed exploration of user engagement across different activities, revealing variations in heart rates, distances, and calorie expenditure. This study contributes unique insights into wearable technology
Bad Each Way Finder
No full textThe aim of this project is to build an application that displaysthe positive expected value each way propositions in UK & Ireland Horse Racing. The expected value is calculated using betting exchange odds in lieu of the probability of winning/losing and comparing these to a sportsbook odds. To complete this an ASP .NET Core back-end API was integrated with the two relevant (Betfair Exchange™ & Sportsbook™) APIs. This data was merged, expected value
was calculated and this was displayed on a user interface via a separate ASP .NET Core application. The application was tested to ensure objectives, requirements were met, and exceptions were handled appropriately.
The resulting application can display live data of positive expected value propositions. Users can track certain propositions of interest. They are also then able to view the history of their tracked propositions to identify trends which may inform their bettin
EV Charging Slot Availability Prediction
Get PDFEV’s are increasing at a very fast rate all around the world and just like petrol pump there is a need for electric travel charging station. The study suggest EV charging slot prediction with the help of machine learning model to predict the charging behavior of electric vehicles. Our study shows the availability of charging station which can help in the optimization of charging for various vehicles and solve problem like optimal charging time, minimum charging cost, optimal duration to charge the vehicle. Machine learning algorithms and data mining techniques are used on a live data set. Study shows the use data imbalanced algorithm SMOTE and with a lot of data cleaning the study shows that KNN model was able to predict the slot availability with maximum accuracy. The models are evaluated on various metrics and KNN shows maximum values in Accuracy, Precison, Recalll and F1-score
Challenges of implementing affiliate marketing in the retail banking sector of Bangladesh
No full textAffiliate marketing offers a low-cost marketing solution for the retail banking industry in Bangladesh, which is forced to face the twin specters of market saturation and high costs resulting from traditional marketing, since the expenses are directly related to performance outcomes. In spite of the rise of internet penetration, the banking industry at the local level has adhered to a great degree to the use of traditional marketing techniques and has lagged behind in implementing affiliate marketing tools, which are digital and scalable in nature. Therefore, the study tried to evaluate the strategic challenges of implementing affiliate marketing and its potential as a core marketing strategy in financial institutions in Bangladesh. This study, through qualitative analysis with the help of in-depth interviews of industry experts, identifies possible benefits and significant barriers of affiliate marketing. Key benefits that have been noted in this regard encompass cost effectiveness and better reach of markets due to the performance-based nature of the tool of affiliate marketing that directly aligns the cost of marketing with outcomes. Technological readiness, regulatory compliance, and customer trust remain the major challenges. It then moves on to examine strategic requirements and needs for the incorporation of affiliate marketing within a broader framework of marketing for achievement of competitive advantages and better market penetration. The study contributes to academic literature and industry practice by offering insight into the intricate process of adoption of digital marketing by the retail banking sector of a developing country like Bangladesh and reporting of strategies that had been custom-designed to overcome specific local challenges
Decoding strategic pathways: Unravelling Amazon Prime Video's key business decision
No full textThis paper is aimed at analyzing Amazon Prime Video’s strategic decisions in relation to those implemented by competitors to differentiate themselves in the SVoD market. The ultimate purpose of the research is to answer, thanks to primary and secondary research, the questions "What is Amazon Prime Video doing and will do to differentiate itself from its competitors and why is it making such moves? Will the changes taking place in this sector lead to a merger of the OTT market with the Pay-TV one?”, and offer three hypothetical scenarios for the future of this market and its players
A thematic analysis to understand the perspectives and the level of support offered to healthcare workers during the pandemic in Ireland
No full textThe COVID-19 pandemic has placed exceptional demands on the healthcare workforce around the world (Vindrola-Padros et al., 2020). The aim of this study was to undertake an exploration of the experiences of healthcare workers (HCWs) who worked during the pandemic and their views about the support they were offered during the pandemic in Dublin, Ireland. Understanding the experiences of HCWs can help identify gaps in healthcare systems and inform efforts to strengthen them.
This primary qualitative study was conducted using semi-structured interviews with six healthcare professionals. Participants were recruited via snowballing sampling technique and included nurses and healthcare assistants. The interview discussion guide consisted of questions on COVID-19-related challenges such as the demands at the workplace, the level of stress and uncertainty to HCWs, availability and quality of personal protective equipment (PPE) and the support as well as coping strategies they received, from the management to handle the pandemic.
Braun and Clarke’s (2021) reflexive thematic analysis generated two themes with eight sub-themes. The two major themes were emotional exhaustion and inconsistent guidelines. The findings from the study indicated that healthcare workers were practising and carrying out duties outside their usual roles and reported very high levels of stress and anxiety. The second theme discusses the lack of consistency, which leads to a number of challenges for HCWs while implementing standardised practices. Understanding the perspectives of healthcare workers would facilitate the hospital administrations as well as managements in Ireland to proactively support healthcare providers during future pandemics by ensuring access to mental health programs, standardising communication and developing plans that will address equipment and supply availability. In addition to this, HCWs are key stakeholders in public health responses to pandemics, and their experiences can inform policies and guidelines related to infection prevention and control, vaccination and other public health interventions, as well as contribute to the development of evidence-based strategies to address pandemics