ARC (Academic Research Collection) (College Dubin)
Not a member yet
359 research outputs found
Sort by
Evaluating the Potential of Ensemble Learning for One Day-Ahead Forecasting of Power System Demand in Ireland
Get PDFAccurate One Day-Ahead Demand Forecasting (ODADF) is crucial for electrical network reliability, the environment, and trading markets. While individual models face challenges in achieving accurate predictions, ensemble learning models have emerged as potential solution. They have achieved success in ODADF in several countries; however, there has been no research conducted for the Irish power system. Therefore, research objectives were formed, to develop a framework of ensemble learning models, evaluate their performance, and examine their potential for ODADF in Ireland, to fill the gap. Experimentation, and CRISP-DM were selected as primary research methodology, and project management framework, respectively. The development of the framework considered a balance between performance and computational complexity of the configurations. Three stacking approaches were considered, such as classifiers and regressors as meta-learners, and heuristic rules. Various potential base-learners were considered, and two methods of supervised problem creation, based on Similar Day (SD) and Moving Window (MW) approaches, were proposed to enhance their pattern recognition in data. The cause-and-effect relationship between ensemble configurations and performance metrics for ODADF in Ireland was established, and the integration method emerged as the primary causal variable. The research methodology was divided into three phases, such as data preparation, experimentation with ensembles architectures, and validation of results. Data preparation included temporal features extraction, Daylight-Saving Time removal, and replacement of missing data and outliers. The results were validated by performance metrics, visual comparison to SDs from neighbouring weeks, and distributions before and after the processing. Investigations into lagged weekly and daily demand, and window size were performed for SD and MW approaches, respectively. Following investigation into correlation between lagged weather variables and demand; temperature, relative humidity and wind speed, lagged by 39-hours were selected as exogenous features. As weather data was distributed locally, three approaches for representative stations were proposed. Scaling of time series, and encoding of temporal features to cyclical and vector formats, were found beneficial to ODADF by correlation study and distributions comparison. Feature selection was performed separately for SD and MW approaches. Given that data from year 2020 was found to be an outlier, datasets were split primarily into training and testing datasets, covering years 2014-2019, and 2021-2022, respectively. Experimentation with base-learners and three integration methods was performed. Training and testing datasets were further split into training and validation subsets, covering years 2014-2018 and 2019, and 2021-2022, respectively. Bayesian optimisation with 10-fold cross-validation was selected for hyperparameters tuning. Potential base-learners were tuned, trained and evaluated on training datasets, and the twenty most promising ones were selected as base-learners. They showed fluctuations in their MAPE across different days of the week, months and hours. Potential ensembles were tuned, trained and evaluated on base-learners’ predictions for years 2015-2019 and training datasets, respectively. In the validation phase, base-learners and classification-based ensembles with hyperparameters inferred from previous phase, were refitted on unseen data, and the base-learners’ predictions for years 2021-2022, respectively, and evaluated on year 2022. The results proved the high potential of classification-based ensembles for ODADF in Ireland. Ensembles of twenty base-learners, with SVM and MLP classifiers as meta-learners, stood out as the most effective solution for ODADF in Ireland. They both achieved the lowest MAPE 1.91%, which was 11.2% improvement in comparison to the best base-learner, SVM (SD) registering MAPE 2.15%. While introduction of SD and MW approaches amplified the diversity of the base-learners’ predictions, incorporating virtual weather stations benefited the performance of classification-based ensembles. They not only harnessed the combined strengths but also mitigated the potential inconsistencies found in individual base-learners, achieving predictions aligned with the distribution of actual demand. Finally, while this research addressed the gap in knowledge, further work, using wider variety of base-learners and their integration methods, is needed to comprehensively bridge this gap
L\u27Oreal
Get PDFThis is a poster presentation focusing on the operational hurdles encountered by the L\u27Oreal make-up brand; encompassing product design intricacies intertwined with environmental concerns
Reinforcement Learning for Stock Option Trading
Get PDFReinforcement learning has recently seen an increase in popularity due to its ability to learn from past experience and its capability of adapting quickly and effectively to new market conditions. This research will focus on reinforcement learning and its importance in trading stock options. Option traders can trade options with one of two option expirations: American or European style. This research will base the analysis on the American expiration style, considered more challenging in trading than the European expiration style. This could lead to the possibility of improving the current trading techniques. In addition, this research aims to understand the role that reinforcement learning plays in trading stock options and evaluate its effectiveness in different market environments. Reinforcement learning has the potential to identify optimal trading strategies for stock options, and could assist current traders in their trading strategies. Trading and markets have existed for millennia, going as far back as Babylon in 2000 BC, with currency exchange and commodities (Kirkpatrick and Dahlquist, 2010). However, markets have evolved and become more complex than in those early trading days. Automation of trading and trading tasks has enabled organisations to act more quickly, consistently, and cost-effectively, all while reducing the risk of human error. The complexity of the markets undeniably increases the difficulty of option trading in dynamic environments. Two questions that arise are: Can Reinforcement Learning models use historical option data to develop effective option trading strategies? Can Reinforcement Learning assist human traders in trading options? These questions are hard to answer at a glance and require robust research and exploration to understand the behaviour of this market segment. Additionally, this research will explore the potential benefits of utilising Reinforcement Learning in stock option trading and how it might be used to modify existing techniques (Moody and Saffell, 2001). The Reinforcement Models that will be explored are Actor-Critical (A2C), Deep Deterministic Policy, Proximal Policy Gradient (DDPG), and Proximal Policy Optimization (PPO). These models use Reinforcement Learning algorithms that train an agent to solve tasks by trial and error. This research will attempt to use these trading agents to develop algorithmic trading 9 strategies, which are difficult for human traders. In chapter 5, there is a complete description of how they work. Ultimately, this research found that Reinforcement Learning can develop trading strategies that could assist human traders. These trading agents are based on machine learning models, which allow them to identify and analyse patterns in the data that human traders may miss. But this research gives evidence to support the results and encourages more work to be done before these can be fully autonomous strategies
Menu Recommendation system using Machine Learning
Get PDFDeveloping a recommendation menu system for restaurants based on the restaurant data and/or city food purchase data to help and change the way restaurants build their menu. Using Data Analysis and Machine Learning to build a project that aims to solve the problem of restaurants and chefs when it comes to preparing menus, the latter with ingredients and dishes that encourage their customers to order more, come back and recommend the restaurant. Helping chefs to create dishes for their restaurants with more accuracy and higher probability to be ordered by their customers. The project will cover tools to build the predictions, the project plan, collect datasets, manipulate data and evaluate the aspects of the situation. The main business goal of our project is to predict what ingredients customers would like to eat and, from that, give restaurants ingredient suggestions to create their next menu. By providing an efficient ingredients decision maker, it will simplify the way menus are elaborated and improve overall customers satisfaction. Another motivating factor in choosing this project was its potential to help society tackle the huge problem of food waste and the inequalities this entails
Version Control Software And Its Possible Impact On Students Academic Integrity
Get PDFEnsuring that students treat their work with integrity has become increasingly difficult in recent years. The advent of Generative AI, Essay Mills, coupled with old fashioned plagiarism and a shift to “online” learning has created a huge shift in the domain of education. Unfortunately, this has manifested as Academic misconduct in many cases and indeed, it could be speculated, that many cases of misconduct are not recognised or discovered. This presentation discusses how version control can be used as a tool to avoid plagiarism
Promoting Student Engagement Through the CCT Student Mentoring Academy
Get PDFhttps://arc.cct.ie/fac_presentations/1013/thumbnail.jp
QA Manual 2022 - 23 Revised
Get PDFThe Quality Assurance Manual for CCT College Dublin, April 2023. Version 5.
Embedding Universal Design for Learning in HECA Colleges: An Exploratory Study
Get PDFThis study explores the strategic implementation of Universal Design for Learning (UDL) in Higher Education Colleges Association (HECA) colleges through an exploratory lens. UDL, rooted in the Centre for Applied Special Technology\u27s (CAST) framework, aims to provide inclusive and challenging learning opportunities for all students. The research investigates the positioning, integration, and impact of UDL within HECA colleges, emphasizing the shift from piecemeal enhancement to a strategic priority. The findings, derived from qualitative data obtained through focus groups, underscore the role of Quality Assurance processes, governance structures, program design, technology, and educational initiatives in advancing UDL. The study contributes to the ongoing dialogue on UDL implementation, suggesting avenues for further research and providing insights for a cohesive, evidence-based approach to UDL within HECA colleges