Combination of Rule-based and Textual Similarity Approaches to Match Financial Entities

Record linkage is a well studied problem [1] with many years of publication history. Nevertheless, there are many challenges remaining to be addressed, such as the topic addressed by FEIII Challenge 2016. Matching financial entities (FEs) is important for many private and governmental organizations. In this paper we describe the problem of matching such FEs across three datasets: FFIEC, LEI and SEC. We were able to achieve an f-measure of 93.78% in the first task, which is comparable to the maximum 97.44%, and 70.44% for the second task, where the maximum is 88.38%.

In Proceedings of the Second International Workshop on Data Science for Macro-Modeling 2016