Swiss National Science Foundation (SNF): Narrative Digital Finance

From EU COST Fin-AI
Jump to navigation Jump to search

Introduction

The project "Narrative Digital Finance: a tale of structural breaks, bubbles & market narratives" aims to delve into the intricate dynamics of financial markets, particularly focusing on the role and impact of different market participants. This project is set to revolutionize the understanding of financial markets by employing advanced quantitative models and frameworks. The project is led by a team of experts, including Prof. Jörg Osterrieder and Prof. Christian Hopp, who bring a wealth of experience in quantitative finance and text analysis, respectively.

The project is unique in its approach to understanding financial markets through the lens of narratives, which have been shown to have a significant impact across various fields such as marketing, journalism, and education. The project aims to extend this understanding to economics, offering a quantitative study of changing popular narratives and their impact on financial markets. The project is distinct in its focus and theoretical foundation, setting it apart from other ongoing research efforts.

The project is expected to have a substantial impact on the field of Finance, particularly in areas related to risk management, asset pricing, and trading strategies. It will employ large-scale financial databases and leverage existing infrastructure for data processing, including dedicated computing resources for deep learning research.

The project is scheduled to commence in July 2023 and has a planned duration of 36 months. It promises to provide scientific innovation in both theory and practice, contributing to a systemic and quantitative understanding that has practical technology applications.

Project Objectives

  • Understanding Financial Markets through Narratives

The primary objective of this groundbreaking project, "Narrative Digital Finance: a tale of structural breaks, bubbles & market narratives," is to revolutionize our understanding of financial markets. We aim to dissect the complex dynamics that govern these markets, focusing particularly on the role and impact of different market participants.

  • Quantitative Modeling and Frameworks

We intend to employ advanced quantitative models and frameworks to achieve this understanding. These models will be based on large-scale financial databases, and we have the infrastructure to handle data processing on a grand scale, including dedicated computing resources for deep learning research.

  • Risk Management and Financial Strategies

The project is expected to have a substantial impact on risk management, asset pricing, and trading strategies. By better understanding the market dynamics, we can offer new techniques and models that are particularly relevant in major fields of quantitative finance.

Scientific Abstract

The project endeavors to augment quantitative-based trading strategies by harnessing the untapped potential of unstructured data, primarily sourced from social media platforms like Twitter and Telegram, as well as financial news outlets. The methodology is bifurcated into two pivotal steps: data collection and data analysis. Initially, textual data from these platforms will be meticulously categorized, dated, and stored in a database managed through SQL solutions. This database will serve as the bedrock for the subsequent analytical phase, which aims to address specific research questions built upon four foundational building blocks. Advanced text analytics and machine learning algorithms will be employed to model this data, thereby providing a fine-grained, real-time information channel that includes both major news stories and minor events. By doing so, the project aspires to offer ex-ante information about market dynamics, thereby filling a critical gap in existing financial decision-making frameworks. The significance of this endeavor lies not only in its potential to disrupt and refine the extant financial services industry but also in its capacity to contribute a nuanced understanding of market dynamics through the analysis of alternative data sources.

Research Plan

Detailed Research Plan Including Methods and Timeline

Phase I: Preliminary Research and Data Collection Review of existing literature on AI in finance and algorithmic trading Data collection from Twitter, Telegram, and financial news outlets Database management through SQL solutions

Phase II: Text Analytics and Machine Learning Development of text analytics algorithms Machine learning model training for market prediction Cross-validation of models

Phase III: Validation and Outcome Preparation Estimation of market prediction models Validation of model robustness Preparation of research outcomes for publication

Timeline: Spanning three years from 2023 to 2026

Expected Outcomes: Original research articles to be prepared in Phase III, presented in workshops and conferences, and submitted for publication in peer-reviewed journals.

Persons

Applicants

  • Joerg Osterrieder, Departement Wirtschaft Berner Fachhochschule BFH, Switzerland
  • Christian Hopp, Departement Wirtschaft Berner Fachhochschule BFH, Switzerland

Principal Investigator

  • Branka Hadji Misheva, Departement Wirtschaft Berner Fachhochschule BFH, Switzerland

Employees

  • Daniel Kucharczyk, Center for Artificial Intelligence ZHAW School of Engineering, Switzerland

Disciplines and keywords

Economics

Conclusion

The project stands at the intersection of artificial intelligence and financial market analysis, aiming to revolutionize the way we understand and predict market dynamics. By integrating various forms of narratives, from text to video, into the analytical framework, the project seeks to provide a more nuanced understanding of financial markets. Advances in data processing and machine learning techniques are harnessed to dissect complex empirical financial relationships that have been lightly studied thus far. The project not only aims to contribute to the academic discourse but also has profound implications for empirical testing in financial markets. It underscores the critical role of narratives in shaping market outcomes and situates these narratives within the audience's social and cultural frameworks. Therefore, the project serves as a pioneering effort in leveraging machine learning techniques to understand how narratives are deeply embedded in the ever-evolving landscape of financial markets.

Links

For more details, visit the Project on SNF webpage.