Bootstrapping a Tagged Corpus through Combination of Existing Heterogeneous Taggers

TL;DR

This paper introduces Combi-bootstrap, a novel method that leverages existing heterogeneous taggers and resources to efficiently annotate corpora with new tagsets, significantly improving accuracy over traditional methods.

Contribution

The paper presents a new approach, Combi-bootstrap, which effectively combines multiple existing taggers and resources for accurate corpus annotation with minimal training data.

Findings

Achieves up to 44.7% error reduction compared to single taggers.

Successfully integrates diverse resources for improved annotation accuracy.

Outperforms ensemble taggers trained on the same small sample.

Abstract

This paper describes a new method, Combi-bootstrap, to exploit existing taggers and lexical resources for the annotation of corpora with new tagsets. Combi-bootstrap uses existing resources as features for a second level machine learning module, that is trained to make the mapping to the new tagset on a very small sample of annotated corpus material. Experiments show that Combi-bootstrap: i) can integrate a wide variety of existing resources, and ii) achieves much higher accuracy (up to 44.7 % error reduction) than both the best single tagger and an ensemble tagger constructed out of the same small training sample.