Henrik Talborn defends his doctoral dissertation Semantic Regression in Finance on April 29.

The dissertation contains three papers:

“Conditionally decomposed sLDA for characteristics decomposition and company classification” presents an extended topic model which address the problem to uncover subtle patterns related to company characteristics in text. 

“A method for semantic regression” proposes a method where interpretable clusters of sentences constitute the approximation of the semantic structure and hence allow for qualitative in-depth analysis of the relation between text and company characteristics. 

“Studying latent risk factors of Fama-French three factor model through semantic regression” address the question whether the factors represent latent systematic risks and studies the underlying logic and mechanisms behind these risk factors. 

HENRIK TALBORN holds a B.Sc in Business Administration, a M.Sc. in Management and Business Strategy and a M.Sc in Engineering Physics from Lund University. He also holds a M.Sc. in Quantitative Finance from ETH Zurich. His main research field is Machine Learning and Artificial Intelligence applied to Finance.