Feature-Rich Named Entity Recognition for Bulgarian Using Conditional Random Fields

TL;DR

This paper introduces a feature-rich Conditional Random Fields approach for Bulgarian Named Entity Recognition, combining language-specific features and resources to achieve high accuracy comparable to English benchmarks.

Contribution

It develops a novel Bulgarian NER system utilizing extensive lexical, syntactic, and morphological features, including rich tagsets and gazetteers, tailored for Bulgarian language processing.

Findings

Achieved F1 score of 89.4% on Bulgarian NER

Effectively integrated language-specific features and unlabeled data

Comparable performance to English NER systems

Abstract

The paper presents a feature-rich approach to the automatic recognition and categorization of named entities (persons, organizations, locations, and miscellaneous) in news text for Bulgarian. We combine well-established features used for other languages with language-specific lexical, syntactic and morphological information. In particular, we make use of the rich tagset annotation of the BulTreeBank (680 morpho-syntactic tags), from which we derive suitable task-specific tagsets (local and nonlocal). We further add domain-specific gazetteers and additional unlabeled data, achieving F1=89.4%, which is comparable to the state-of-the-art results for English.