QuantfolioX: portfolio management application using large language model technology

Abstract

This final year project report conducts a thorough analysis of the limitations inherent in Traditional Portfolio Management and existing Robo-advisor models. Emphasis is placed on critical aspects such as portfolio monitoring, construction, and recommendation, with a particular focus on integrating Large Language Model (LLM) technology to address identified shortcomings. The report begins with an in-depth exploration of the drawbacks associated with Traditional Portfolio Management methodologies and the current state of Robo-advisors, laying the groundwork for understanding the motivations behind proposed enhancements. The proposed solution, QuantfolioX, is a web application for portfolio management. A significant contribution of this project is the innovative use of LLM technology to improve user interactions and enhance the explainability of portfolios. LLM serves a dual role in enhancing the live interaction interface with users and contextualizing portfolio recommendations within the current market environment. Furthermore, a novel aspect of this application is the integration of a ML-Driven Approach in portfolio allocation, aiming to address inadequacies observed in traditional portfolio management techniques. This involves adapting dynamically to diverse market conditions and providing a more responsive investment strategy.