Die Dissertation befasst sich mit der Anwendung von statistischem und maschinellem Lernen zur Modellierung der Verlustquote bei Ausfall (LGD). Im Forschungsgebiet der LGD-Modellierung gibt es eine Reihe von Fragen und Problemen, die bisher in der Literatur nicht berücksichtigt wurden. Erstens ist unklar, welche Merkmale einer LGD-Verteilung für die Prognosefähigkeit von Schätzmethoden entscheidend sind und welche Schätzmethode für die LGD-Modellierung am besten geeignet ist. Zweitens besteht ein Zielkonflikt zwischen der Transparenz und der Prognosegenauigkeit bei LGD-Schätzmethoden. Komplexe maschinelle Lernalgorithmen weisen eine bessere Vorhersageleistung auf, allerdings auf Kosten einer geringeren Erklärbarkeit. Umgekehrt bietet die lineare Regression eine hohe Interpretierbarkeit, scheint aber eine geringere Prognosegenauigkeit aufzuweisen. Um diesen Zielkonflikt zu lösen, besteht ein geeigneter Ansatz darin, die Vorhersagegenauigkeit der interpretierbaren linearen Regression durch maschinelles Lernen zu verbessern. Drittens stellt die Selektion optimaler Clustervariablen in der gruppierten Modellierung eine zu lösende Herausforderung dar. Die offenen Forschungsfragen werden in der Dissertation anhand von Kreditausfalldaten der Global Credit Data empirisch beantwortet.

This book is about making machine learning models and their decisions interpretable. After exploring the concepts of interpretability, you will learn about simple, interpretable models such as decision trees, decision rules and linear regression. Later chapters focus on general model-agnostic methods for interpreting black box models like feature importance and accumulated local effects and explaining individual predictions with Shapley values and LIME. All interpretation methods are explained in depth and discussed critically. How do they work under the hood? What are their strengths and weaknesses? How can their outputs be interpreted? This book will enable you to select and correctly apply the interpretation method that is most suitable for your machine learning project.

The long-awaited, comprehensive guide to practical credit risk modeling Credit Risk Analytics provides a targeted training guide for risk managers looking to efficiently build or validate in-house models for credit risk management. Combining theory with practice, this book walks you through the fundamentals of credit risk management and shows you how to implement these concepts using the SAS credit risk management program, with helpful code provided. Coverage includes data analysis and preprocessing, credit scoring; PD and LGD estimation and forecasting, low default portfolios, correlation modeling and estimation, validation, implementation of prudential regulation, stress testing of existing modeling concepts, and more, to provide a one-stop tutorial and reference for credit risk analytics. The companion website offers examples of both real and simulated credit portfolio data to help you more easily implement the concepts discussed, and the expert author team provides practical insight on this real-world intersection of finance, statistics, and analytics. SAS is the preferred software for credit risk modeling due to its functionality and ability to process large amounts of data. This book shows you how to exploit the capabilities of this high-powered package to create clean, accurate credit risk management models. Understand the general concepts of credit risk management Validate and stress-test existing models Access working examples based on both real and simulated data Learn useful code for implementing and validating models in SAS Despite the high demand for in-house models, there is little comprehensive training available; practitioners are left to comb through piece-meal resources, executive training courses, and consultancies to cobble together the information they need. This book ends the search by providing a comprehensive, focused resource backed by expert guidance. Credit Risk Analytics is the reference every risk manager needs to streamline the modeling process.

Customer survey studies deals with customers, consumers and user satisfaction from a product or service. In practice, many of the customer surveys conducted by business and industry are analyzed in a very simple way, without using models or statistical methods. Typical reports include descriptive statistics and basic graphical displays. As demonstrated in this book, integrating such basic analysis with more advanced tools, provides insights on non-obvious patterns and important relationships between the survey variables. This knowledge can significantly affect the conclusions derived from a survey. Key features: Provides an integrated, case-studies based approach to analysing customer survey data. Presents a general introduction to customer surveys, within an organization’s business cycle. Contains classical techniques with modern and non standard tools. Focuses on probabilistic techniques from the area of statistics/data analysis and covers all major recent developments. Accompanied by a supporting website containing datasets and R scripts. Customer survey specialists, quality managers and market researchers will benefit from this book as well as specialists in marketing, data mining and business intelligence fields.

A risk measurement and management framework that takes model risk seriously Most financial risk models assume the future will look like the past, but effective risk management depends on identifying fundamental changes in the marketplace as they occur. Bayesian Risk Management details a more flexible approach to risk management, and provides tools to measure financial risk in a dynamic market environment. This book opens discussion about uncertainty in model parameters, model specifications, and model-driven forecasts in a way that standard statistical risk measurement does not. And unlike current machine learning-based methods, the framework presented here allows you to measure risk in a fully-Bayesian setting without losing the structure afforded by parametric risk and asset-pricing models. Recognize the assumptions embodied in classical statistics Quantify model risk along multiple dimensions without backtesting Model time series without assuming stationarity Estimate state-space time series models online with simulation methods Uncover uncertainty in workhorse risk and asset-pricing models Embed Bayesian thinking about risk within a complex organization Ignoring uncertainty in risk modeling creates an illusion of mastery and fosters erroneous decision-making. Firms who ignore the many dimensions of model risk measure too little risk, and end up taking on too much. Bayesian Risk Management provides a roadmap to better risk management through more circumspect measurement, with comprehensive treatment of model uncertainty.

This paper discusses the impact of the rapid adoption of artificial intelligence (AI) and machine learning (ML) in the financial sector. It highlights the benefits these technologies bring in terms of financial deepening and efficiency, while raising concerns about its potential in widening the digital divide between advanced and developing economies. The paper advances the discussion on the impact of this technology by distilling and categorizing the unique risks that it could pose to the integrity and stability of the financial system, policy challenges, and potential regulatory approaches. The evolving nature of this technology and its application in finance means that the full extent of its strengths and weaknesses is yet to be fully understood. Given the risk of unexpected pitfalls, countries will need to strengthen prudential oversight.