Saudi Cultural Missions Theses & Dissertations
Permanent URI for this communityhttps://drepo.sdl.edu.sa/handle/20.500.14154/10
Browse
3 results
Search Results
Item Restricted Deep Learning Approaches for Multivariate Time Series: Advances in Feature Selection, Classification, and Forecasting(New Mexico State University, 2024) Alshammari, Khaznah Raghyan; Tran, Son; Hamdi, Shah MuhammadIn this work, we present the latest developments and advancements in the machine learning-based prediction and feature selection of multivariate time series (MVTS) data. MVTS data, which involves multiple interrelated time series, presents significant challenges due to its high dimensionality, complex temporal dependencies, and inter-variable relationships. These challenges are critical in domains such as space weather prediction, environmental monitoring, healthcare, sensor networks, and finance. Our research addresses these challenges by developing and implementing advanced machine-learning algorithms specifically designed for MVTS data. We introduce innovative methodologies that focus on three key areas: feature selection, classification, and forecasting. Our contributions include the development of deep learning models, such as Long Short-Term Memory (LSTM) networks and Transformer-based architectures, which are optimized to capture and model complex temporal and inter-parameter dependencies in MVTS data. Additionally, we propose a novel feature selection framework that gradually identifies the most relevant variables, enhancing model interpretability and predictive accuracy. Through extensive experimentation and validation, we demonstrate the superior performance of our approaches compared to existing methods. The results highlight the practical applicability of our solutions, providing valuable tools and insights for researchers and practitioners working with high-dimensional time series data. This work advances the state of the art in MVTS analysis, offering robust methodologies that address both theoretical and practical challenges in this field.38 0Item Restricted Essays on Risk Management and Portfolio Allocation Using Tail Measures(University of Southampton, 2024-07-09) Alruwaili, Faridah; Olmo, Jose; Xu, Huifu; Lu, ZudiThis thesis delves into effective methods for managing risks and decision-making processes in finance. The research comprises three main chapters, each addressing critical challenges in time series analysis, portfolio optimization, and risk assessment. Chapter 2 introduces the Robust Model Averaging Marginal Regressions (RMAMAR) procedure, a novel approach that combines one-dimensional marginal regression functions to approximate conditional regression functions robustly. By employing local linear estimation and robust M-estimators, RMAMAR addresses the curse of dimensionality and enhances parameter estimation accuracy, particularly in high-dimensional datasets. Chapter 3 extends dynamic portfolio choice methodologies under Expected Utility (EU) frameworks to incorporate investors’ quantile preferences, focusing on specific quantiles of the returns distribution. Through empirical applications and simulations, this chapter demonstrates the effectiveness of constructing optimal portfolios under quantile preferences with multiple conditioning variables, showcasing superior performance during market crises. Chapter 4 proposes a new approach to backtesting risk measures by introducing a univariate score function that combines the marginal/conditional score functions for forecasting Value-at-Risk (VaR) and Systemic Risk (SR). This method ensures an equitable and comprehensive assessment of both risk measures, overcoming the limitations of existing methods that prioritize one measure over the other based on the equality of VaR across models. Furthermore, Chapter 4 conducts a comparative analysis of different identification functions for backtesting, including the one-dimensional function proposed by Banulescu-Radu et al. (2021) and the two-dimensional function introduced byFissler and Hoga (2023), to evaluate the potential risk associated with employing identification functions that are not strictly defined for backtesting purposes. Overall, this thesis contributes to advancing risk management and decision-making methodologies by providing robust and practical strategies.25 0Item Restricted Forecasting Risk and Failure in the Hedge Fund Industry(University of East Anglia, 2024-06-30) Aldhahi, Huda Ibrahim Saad; Kourtis, Apostolos; Markellos, RaphaelThis dissertation aims to provide a comprehensive understanding of hedge fund performance, volatility forecasting, and survival analysis based on three extensive studies. It aims to extensively evaluate the performance of various hedge fund indices and examine the factors influencing hedge fund survival using an extensive dataset between 1994-2020. This thesis is constructed in three parts. In the first study, conduct a comprehensive analysis of prominent volatility forecasting models applied to different hedge fund indices and time horizons. The results indicate asymmetric EGARCH and TGARCH models as optimal choices for forecasting daily and weekly hedge fund volatility. Moreover, the study identifies IGARCH and LRE models as inferior alternatives across all indices and horizons examined. The second study deeply investigates the survival of hedge funds by exploring critical factors behind their failures using survival analysis techniques such as non-parametric survival analysis, Semi-parametric Cox proportional hazard, and Weibull AFT methods. This research reveals age, size, and performance as critical determinants for hedge funds' longevity. Conversely, volatility, advanced notice period, and efficiency values negatively affect hedge fund survival. The relationship between management fees, leverage employed, lockup periods, and fund survival rates exhibit mixed results based on measurements, fund styles, and evaluation periods studied. The third study evaluates hedge fund performance through data envelopment analysis (DEA) to provide an accurate ranking of different performances. The findings offer insights into the instability of various hedge fund strategies in diverse time horizons. Additionally, it examines the impact of major economic crises on the performance of hedge funds. Ultimately, this research contributes significantly to investors' and fund managers' understanding by identifying high-performing funds to optimize portfolio diversification effectively. The overarching objective of this dissertation is to provide investors and fund managers with a comprehensive and detailed understanding of hedge funds by investigating various aspects such as volatility forecasting techniques, key drivers of longevity, and precise performance measurement using data envelopment analysis. By examining these critical elements with extensive datasets and innovative methodologies, this dissertation aims to contribute significantly to the existing literature in the field while providing valuable guidance for investment decisions and portfolio diversification strategies.17 0