Applications of Artificial Intelligence to Credit Scoring

In recent years, Artificial Intelligence has made important developments, resulting in its use in several economic fields. In particular, the application of artificial intelligence has led to better performance of credit scoring models. The thesis will look at the innovations that these new techniques can produce to the economy in terms of extension of the credit to individuals or SME previously rejected, the use of new non-traditional data in order to evaluate credit risk, and the approach of the authorities in regulating these new issues. The first chapter provides an overview of the artificial intelligence applications to the economy, understanding machine learning methods and which are the context where it is optimal to use them. The second chapter is centred on the use of AI in relation to credit scoring, analysing what are the implications of this phenomena, why these methods have an advantage over traditional methods, taking into account the specific economic phase. The third chapter focuses on empirical data, and compares the results of using machine learning with those of traditional approaches, in their ability to measure credit scoring. The findings of this thesis highlight, in different contexts, the benefits that AI models produce on classification accuracy and performance when compared to other credit scoring methods.