A Named Entity Recognition System for Dutch

Abstract

We describe a Named Entity Recognition system for Dutch that combines gazetteers, handcrafted rules, and machine learning on the basis of seed material. We used gazetteers and a corpus to construct training material for Ripper, a rule learner. Instead of using Ripper to train a complete system, we used many different runs of Ripper in order to derive rules which we then interpreted and implemented in our own, hand-crafted system. This speeded up the building of a hand-crafted system, and allowed us to use many different rule sets in order to improve performance. We discuss the advantages of using machine learning software as a tool in knowledge acquisition, and evaluate the resulting system for Dutch.

Character Profiling in Low-Resource Language Documents

“…This research question is especially under-investigated for the lit-erary domain. In the Namescape project, named entity recognizers were trained on Dutch literary texts to recognize names in Dutch fiction [10]. Grammatical structure was exploited in recognizing proper names in French novels [5].…”

“…Specifically, we investigate the impact of human annotation for character profiling in literary text written in low-resource lan-guages. Automatic profiling involves NLP tasks such as named entity recognition (NER) [10] and syntactic analysis [5]. Although syntactic features have been shown to be beneficial for information extraction [5,21,34], off-the-shelf parsers are rarely available for…”

Diversity Checker

“…Fortunately, NER is one of the NLP tasks that has seen results published for many languages, not just English [25]. Besides tools that work on multiple languages [1,19,29], there are things specifically focused on Dutch [7,8].…”

“…Frequently, NER tools have classified entities in very coarse classes of entities, usually including person, organization and location, e.g. [7]. Attempts are being made to classify more fine-grained categories, especially for the person category, e.g.…”