MBA (MIS) BUIC

Benchmarking LLM Models Against Traditional Machine Learning and Deep Learning Models for Stock Prediction

Mon, 01 Jan 2024 00:00:00 GMT

Benchmarking LLM Models Against Traditional Machine Learning and Deep Learning Models for Stock Prediction Muhammad Abubakar Nawaz, 01-321232-018 This thesis investigates the application of Large Language Models (LLMs) to stock price prediction, a challenging task traditionally addressed through technical analysis and, more recently, machine learning (ML) and deep learning (DL) techniques. The research benchmarks the performance of an LLM, based on Google's Gemini architecture, against established ML models (XGBoost, Random Forest) and a DL model (Gated Recurrent Unit, GRU) in predicting the binary direction (up or down) of Apple Inc. (AAPL) stock's daily closing price. The models were trained and evaluated using five years of historical price and volume data from Yahoo Finance, transformed into five widely used technical indicators: Exponential Moving Average (EMA), Relative Strength Index (RSI), Triple Exponential Moving Average (TEMA), Chaikin Money Flow (CMF), and Money Flow Index (MFI). These indicators served as input features, with sequences of 20 days used for the LLM and GRU. The study found that the LLM significantly outperformed both XGBoost and Random Forest in terms of accuracy, F1-score, and AUC-ROC (p < 0.05), while achieving results comparable to the GRU, a model specifically designed for sequential data. These findings support the hypothesis that LLMs, with their Transformer-based architecture and attention mechanism, are particularly well-suited to capturing long-range dependencies and learning complex patterns within sequences of technical indicators. The research contributes to the nascent field of LLMs in finance by demonstrating their potential to enhance and potentially transform traditional technical analysis. The results suggest that LLMs can offer a significant advantage in stock prediction accuracy, potentially leading to more informed investment decisions. Further research is recommended to explore the generalizability of these findings to other stocks and market conditions, incorporate additional data sources, optimize LLM architectures for this task, and develop methods for interpreting the models' decision-making processes. The results of applying LLMs on technical indicators will provide more insights whether these models are able to predict the stock price accurately or not. It will also provide insights whether these models are better than the traditional models or not. Also, the results indicate that deep learning models are more suitable for financial forecasting. Supervised by Mr. Jawwad Ijaz

Mitigating Corruption with the Use of Effective Financial Information Systems

Mon, 01 Jan 2024 00:00:00 GMT

Mitigating Corruption with the Use of Effective Financial Information Systems Usama Zahid, 01-221232-018 Corruption is a pervasive issue that significantly undermines economic development, governance, and public trust globally. Traditional anti-corruption methods have often fallen short due to their inability to address systemic inefficiencies and human-driven malpractices. The advent of Financial Information Systems (FIS) presents a transformative approach to mitigating corruption by leveraging technology to enhance transparency, accountability, and efficiency in financial operations. This thesis examines the impact of FIS on corruption mitigation through an extensive review of secondary data, including case studies, government reports, and academic literature. Key findings highlight that FIS, such as Enterprise Resource Planning (ERP) systems, e-procurement platforms, and digital tax administration systems, significantly reduce corruption risks by automating processes, providing real-time monitoring, and improving financial traceability. For instance, countries adopting e-procurement systems have seen marked reductions in public sector corruption due to increased competition and transparency. Similarly, the integration of blockchain technology in financial systems has bolstered data integrity and minimized fraudulent activities. However, challenges such as high implementation costs, resistance to change, and cybersecurity threats pose significant barriers to the widespread adoption of FIS. This study underscores the importance of robust legal frameworks, stakeholder engagement, and capacity building in ensuring the successful implementation of FIS. It also highlights the potential of emerging technologies, including artificial intelligence and blockchain, to further enhance the effectiveness of financial information systems in curbing corruption. The findings provide actionable recommendations for policymakers, businesses, and governments, advocating for a strategic approach to integrating FIS into anti-corruption frameworks. By doing so, societies can move closer to achieving greater economic equity and institutional integrity. Supervised by Mr. Qazi Haseeb Yousaf

Impact of Instrumental and Non-Instrumental Gratification on ChatGPT Usage at the Workplace

Mon, 01 Jan 2024 00:00:00 GMT

Impact of Instrumental and Non-Instrumental Gratification on ChatGPT Usage at the Workplace Syeda Bint-e-Huda Kazmi, 01-321232-045 This study examines the impact of instrumental and non-instrumental gratifications on the usage of ChatGPT in workplace settings, focusing on the factors that influence employees' intention to adopt this conversational AI tool. Leveraging the Uses and Gratifications (U&G) theory alongside the Technology Acceptance Model (TAM), this research explores how task-oriented benefits such as productivity and efficiency (instrumental gratification), enjoyment, and personal innovativeness (non-instrumental gratification) drive adoption behaviours. The study also investigates the role of gain with user tendency, which includes satisfaction and individual impact, in shaping the intention to use ChatGPT at work. A quantitative research approach was adopted, with data collected through a structured questionnaire distributed to 200 employees from diverse workplaces. The questionnaire included Likert-scale items to measure perceptions of ChatGPT's utility, ease of use, emotional engagement, and broader impacts on professional performance. The collected data were analyzed using SPSS software to identify relationships between the independent variables, instrumental gratification, non-instrumental gratification, and gain with user tendency and the dependent variable, intention to use ChatGPT. Findings indicate that instrumental and non-instrumental gratifications significantly influence employees' intention to use ChatGPT. Moreover, gain with user tendency is critical, enhancing the relationship between gratifications and usage intention. This research contributes to the literature on AI adoption in professional settings by providing a complete understanding of the dual-purpose motivations driving ChatGPT usage. Supervised by Mr. Jawwad Ijaz

The Effect of AI Powered Sentiment Analysis on Consumer Complaint Resolution in Ecommerce: A Comparative Study of Human Vs AI Meditation Support

Mon, 01 Jan 2024 00:00:00 GMT

The Effect of AI Powered Sentiment Analysis on Consumer Complaint Resolution in Ecommerce: A Comparative Study of Human Vs AI Meditation Support Muhammad Ali Malik, 01-321232-019 Artificial intelligence's (AI) quick development has revolutionized e-commerce and improved customer service effectiveness. With an emphasis on resolution speed, emotional understanding, customer happiness, and cost-efficiency, this study compares the effectiveness of traditional human mediation with AI-powered sentiment analysis in handling customer complaints. Quantitative data from online surveys was analyzed using a mixed-methods approach and statistical tools such as the Mann-Whitney U Test, ANOVA, and Chi-Square Test. This investigation provided valuable insights into the advantages and disadvantages of both approaches. The literature review emphasizes the importance of human mediators, the development of technology for emotion analysis, and theoretical frameworks such as the Emotional Contagion Theory and the Acceptance Model of technologies. The results show that AI is highly effective and scalable, and it drastically cuts down on the amount of time needed to process complaints (Iqbal et al., 2023). The significance of human mediators in compassionate consumer encounters is highlighted by their remarkable ability to manage intricate emotional nuances (Banga and Peddireddy, 2023). This study proposes a hybrid AI-human mediation framework that strikes a compromise between emotional intelligence and technology efficacy in order to optimize complaint settlement processes (Sai Mounika Inavolu, 2024). Furthermore, discussed are ethical issues like data security and AI transparency. By providing practical guidance for integrating AI and human approaches in e-commerce, the study broadens the body of existing literature and opens the door for further investigation into collaborative mediation models. By boosting operational efficiency and preserving client trust, these models support service delivery and client pleasure. Supervised by Mr. Jawwad Ijaz