Deep Learning Applications in Enterprise Data Science

September 2018 (completed) Preproposal PhD-Thesis

Project Description

Deep Neural Networks (DNNs) are a powerful machine learning tool, loosely inspired by the structure of the human brain. Improvements in the field of DNNs, combined with an increase in computational power and available data, played an essential part in the recent rise of artificial intelligence applications. So far, the impact of deep learning has been most prevalent in a few specific areas, like image recognition, text- and speech processing. However, DNNs ability to handle large amounts of structured as well as unstructured data give them a considerable potential to create new value adding solutions in data analytics for enterprises. Big data initiatives and the rise of mobile and internet of things technologies leave companies with an enormous amount of raw data. DNNs seem to be the ideal technology to turn these data into useful knowledge.

This dissertation project uses action design research to explores the potential and challenges of applying deep learning methods for data analytics in enterprises. It will explore how deep learning methods can create new ways of analyzing enterprise data and how they can lead to value adding applications that make use of these data.

Keywords

Data analytics Machine Learning Neural Networks

Project Participants

Dr. rer. oec. Marcus Basalla M.Sc.

- PhD-Student

PhD-Student

Prof. Dr. Jan vom Brocke

- Supervisor

Visiting Professor - Information Systems and Process Science

Supervisor

Deep and (Un-) Constrained Portfolio Optimization

December 2023 until June 2024 (completed) FFF-Funding Project

Project Description

Since its birth in the 1950ies, portfolio optimization has suffered from errors regarding the esti-mation of the input parameters (Michaud, 1989). To overcome the resulting underperformance, recent advances in Machine Learning mitigate the impact of estimation errors by directly opti-mizing portfolio weights from raw input data, e.g., using deep neural networks. However, these initial approaches still lack one important practical aspect by neglecting the (portfolio) weight constraints faced by real world asset management companies (e.g., short sale restrictions, in-dustry exposure limitation, factor exposure targets, diversification requirements, or upper bounds on transaction costs). We strive to improve on existing approaches by allowing for the implementation of such constraints. At the conclusion of this project, in addition to a scientific paper, we plan to provide a software toolbox in R and/or Python that implements our findings.

Participating Institutions

Financial Economics / Project Lead

Project Participants

Assoz. Prof. Dr. Sebastian Stöckl

- Principal Investigator

Associate Professor - Financial Economics Academic Director - Liechtenstein Undergraduate and Graduate School Official Representative AACSB - University of Liechtenstein

Principal Investigator

Decision-making in crowdfunding

September 2017 until August 2020 (completed) PhD-Thesis

Project Description

Crowdfunding is an Internet-based approach to raising capital through collective efforts of many individuals. In recent years, people have created tens of thousands of projects and campaigns that have collected billions of dollars through crowdfunding. Four basic crowdfunding practices have emerged: (1) donation-based crowdfunding, whereby investors are not compensated for their funding and which is usually used for charity projects; (2) lending-based crowdfunding, whereby investors are compensated with interest and which is usually used for private loans; (3) equity-based crowdfunding, whereby investors are compensated with shares or dividends and which is usually used for start-ups; and (4) reward-based crowdfunding, whereby investors are compensated with so-called rewards and which is usually used for creative projects. As of recently, established companies have also developed an interest in crowdfunding, although they typically do not use it to collect money but for purposes such as marketing, open innovation, and prototyping.

Since its fundamental concepts, techniques, and practices are constantly changing, researchers from various fields increasingly study crowdfunding, including Information Systems researchers. However, few researchers have studied decision-making at the individual level with the help of experiments. Against this backdrop, the dissertation project experimentally explores decision processes in crowdfunding through a behavioral-economics lens. The dissertation is paper-based and thus covers a series of studies, each of which examines individuals' crowdfunding decisions from a different perspective.

Keywords

Crowdfunding Behavioural Finance Experimental Research

Participating Institutions

Hilti Chair of Business Process Management / Doctoral instiution

Project Participants

Dr. rer. oec. Lena Franziska Kaiser

- PhD-Student

PhD-Student

Prof. Dr. Jan vom Brocke

- Supervisor

Visiting Professor - Information Systems and Process Science

Supervisor

Prof. Dr. Alexander Mädche

- Co-Supervisor

Co-Supervisor

Decision-making in crowdfunding

September 2017 (completed) Preproposal PhD-Thesis

Project Description

Keywords

Crowdfunding Behavioural Finance Experimental Research

Project Participants

Prof. Dr. Jan vom Brocke

- Supervisor

Visiting Professor - Information Systems and Process Science

Supervisor

Dr. rer. oec. Lena Franziska Kaiser

- PhD-Student

PhD-Student

Decision methods in pension finance: Large-scale optimization

January 2022 until December 2022 (completed) FFF-Funding Project

Project Description

In a former project funded by the FFF (Fin-19-2) we have developed a software package that models the financial decisions an individual insured in Liechtenstein has to take regarding his/her pensions. This includes optimal consumption and saving decisions in the context of the specific financial situation of the insured. It also includes the allocation of savings to the three pillars as well as the direct investment in various financial assets and decisions on whether to receive the entire savings as a lump sum when entering retirement or the option to transform the savings into a (life-long) annuity.
As those optimizations require a large amount of time (10 minutes) per parameter setting we have run a limited number of such optimizations (approx. 900’000) for a pre-specified parame-ter grid (amounting to 3.2 mn combinations) and then used the optimizations to train a Machine Learning model to accurately and quickly (2-5 seconds) approximate these optimal decisions. Even after the end of the former project, we continued running these optimizations to improve the accuracy of the Machine Learning (ML) model, but due to the Ransomware attack at the University of Liechtenstein, many of our results as well as the computing resources were lost and are not recoverable. In this project, we therefore apply for additional funding to finish the computations on a large-scale optimization cluster, retrain our ML model and then provide the insured with accurate and quickly available optimal pension decisions on the dedicated website (app currently unavailable due to the server loss): https://apps.resqfin.com/pfli.

Project Participants

Assoz. Prof. Dr. Sebastian Stöckl

- Principal Investigator

Associate Professor - Financial Economics Academic Director - Liechtenstein Undergraduate and Graduate School Official Representative AACSB - University of Liechtenstein

Principal Investigator

Decision methods and tools in the context of pension finance

September 2019 until August 2020 (completed) FFF-Funding Project

Project Description

In this project we developed an R-package (available through github at https://github.com/sstoeckl/pensionfinanceLi) to optimize decisions individuals in Liechtenstein's pension system have to take. The package contains several optimizers as well as a documentation (available through 'vignette("model")' once the package is installed). We have started the optimization for a feasible parameter grid to determine which variables are the most relevant drivers of optimal pension decisions. Based on the results we have trained three machine learning models (a hyper parameter-tuned random forest performs best) to allow individuals to receive faster and near-optimal decisions without having to wait for the individual optimization on each run (up to 25 minutes on a regular CPU). Predictions from these models are available to the public at https://apps.resqfin.com/pfli where - based on each persons individual settings.

Project results:

Participating Institutions

Forschungsförderungsfonds der Universität Liechtenstein / Sponsor Chair in Finance / Project Lead

Project Participants

Assoz. Prof. Dr. Sebastian Stöckl

- Principal Investigator

Associate Professor - Financial Economics Academic Director - Liechtenstein Undergraduate and Graduate School Official Representative AACSB - University of Liechtenstein

Principal Investigator

Prof. Dr. Michael Hanke

- Contributor

Professor - Finance Dean - Liechtenstein Business School

Contributor

Decision Criteria for the Domiciling of Investment Funds

September 2005 until March 2006 (completed) KTT

Project Description

The project addresses to the study of Liechtenstein's fund location and its competitiveness in international comparison. The specific topics are sub classified into three phases. Phase 1 investigates the attractiveness of Liechtenstein as a fund domicile. It is represented by the evaluation of a survey among domestic and foreign German-speaking fund leaderships and independent asset managers from abroad. Phase 2 deals with the segment of alternative investments. Liechtenstein's legal framework for qualified investors and alternative investment funds based on the new law of investment undertakings (IUG) are opposed to comparable competitor's products of other fund locations. The closing part (phase 3) concentrates on the greenpaper of the European Commission and its potential ascendancies on Liechtenstein by mean of reactions of institutions, investors and public authorities of the European fund industry.

Participating Institutions

Chair in Business Administration, Banking and Financial Management / Project Lead

Project Participants

Prof. em. Dr. Marco J. Menichetti

- Principal Investigator

Professor Emeritus - Liechtenstein Business School

Principal Investigator

Lic. oec. HSG Oliver C. Oehri

- Project Collaborator

Project Collaborator

MMag. Dr. Wilfried Amann

- Project Collaborator

Project Collaborator

Dr. Marcel Vaschauner MBA

- Project Collaborator

Project Collaborator

Data driven company valuation and IPO performance prediction

April 2018 until March 2020 (completed) FFF-Funding Project

Project Description

An important part of every initial public offering (IPO) is an accurate valuation of the company planning to sell shares. Most established valuation methods only take hard financial facts into account. However soft facts, like management experience, were shown to play an important role in the success of companies, especially for start-ups and SMEs. In practice it is common to use a mix of several valuation models, which makes the process less transparent for investors.
This project aims to develop a company valuation model that makes the valuation process more efficient as well as more accurate and transparent for both investors and companies. The project evaluates existing valuation models and explores technologies to integrate further information into the valuation process like consumers' opinion on the company, the experience of the managerial staff, expert opinions as well as a more accurate prediction of the company's future performance. To this end state of the art algorithms in the fields of text-mining, sentiment analysis, machine learning and crowdsourcing are investigated. These methods will be used to facilitate the automatic extraction of financial data from existing records, analyze a company's public reputation from (social) media data and predict its future performance.
By partnering with Own a startup company aiming to disrupt the equity market with a block chain based sales platform for company shares, we ensure that our work will have a direct impact on the equity market. This cooperation further allows us to profit from market insight and an environment to evaluate the use of our model.

Keywords

Data analytics

Participating Institutions

Forschungsförderungsfonds der Universität Liechtenstein / Sponsor Hilti Chair of Business Process Management / Project Lead

Project Participants

Dr. rer. oec. Lena Franziska Kaiser

- Project Collaborator

Project Collaborator

Prof. Dr. Johannes Schneider

- Project Collaborator

Professor - Artificial Intelligence and Data Science

Project Collaborator

Prof. Dr. Jan vom Brocke

- Professor

Visiting Professor - Information Systems and Process Science

Professor

Dr. rer. oec. Marcus Basalla M.Sc.

- Principal Investigator

Principal Investigator

Publications

Data governance in financial services companies

September 2017 until July 2022 (completed) PhD-Thesis

Project Description

Big Data Analytics (BDA) refers to analytical techniques applied to large and complex data sets. Furthermore, BDA encompasses advanced data storage, management, analysis, and visualization technologies. Application areas comprise text analytics, audio analytics, video analytics, and predictive analytics. In recent years, BDA has quietly descended on many industries, from e-commerce and manufacturing to the health sector and the financial services industry. Particularly financial services companies have commenced several BDA initiatives including the automated supervision of client portfolios, trade recommendation systems, and robo-advisory. While many recipes for the success of BDA initiatives center around building BDA capabilities, the implementation of BDA requires further accompanying structural, relational, and process changes to preserve organizational success over time. Hence, the thesis aims to explore those required organizational changes and targets to define a BDA governance model. In doing so, the thesis focuses on investigating BDA initiatives within the financial services industry.