Domain Adaption of Named Entity Recognition to Support Credit Risk Assessment

Risk assessment is a crucial activity for financial institutions because it helps them to determine the amount of capital they should hold to assure their stability. Flawed risk assessment models could return erroneous results that trigger a misuse of capital by banks and in the worst case, their collapse. Robust models need large amounts of data to return accurate predictions, the source of which is text-based financial documents. Currently, bank staff extract the relevant data by hand, but the task is expensive and timeconsuming. This paper explores a machine learning approach for information extraction of credit risk attributes from financial documents, modelling the task as a named-entity recognition problem. Generally, statistical approaches require labelled data for learn the models, however the annotation task is expensive and tedious. We propose a solution for domain adaption for NER based on out-of-domain data, coupled with a small amount of in-domain data. We also developed a financial NER dataset from publicly-available financial documents.