DErivBase: Inducing and Evaluating a Derivational Morphology Resource for German
Article Structure
Abstract
Derivational models are still an under-researched area in computational morphology.
Introduction
Morphological processing is generally recognized as an important step for many NLP tasks.
Related Work
Computational models of morphology have a long tradition.
Framework
In this section, we describe our rule-based model of derivation, its operation to define derivational families, and the application of the model to German.
Building the Resource
4.1 Derivational Rules
Evaluation
5.1 Baselines
Results
6.1 Quantitative Evaluation
Conclusion and Future Work
In this paper, we present DERIVBASE, a derivational resource for German based on a rule-based framework.
Topics
precision and recall
Appears in 10 sentences as: Precision and recall (2) precision and recall (8)
In DErivBase: Inducing and Evaluating a Derivational Morphology Resource for German
- We conduct a thorough evaluation of the induced derivational families both regarding precision and recall .Page 2, “Introduction”
- We then revised the rules with the aim of increasing both precision and recall .Page 4, “Building the Resource”
- To obtain reliable estimates of both precision and recall , we decided to draw two different samples: (l) a sample of lemma pairs sampled from the induced derivational families, on which we estimate precision (P-sample) and (2) a sample of lemma pairs sampled from the set of possibly derivationally related lemma pairs, on which we estimate recall (R-sample).Page 6, “Evaluation”
- Table 5: Precision and recall on test samplesPage 7, “Evaluation”
- We omit the F1 score because its use for precision and recall estimates from different samples is unclear.Page 7, “Results”
- The string distance-based approaches achieve more balanced precision and recall scores.Page 7, “Results”
- Note that for these methods, precision and recall can be traded off against each other by varying the number of clusters; we chose the number of clusters by optimizing the F1 score on the calibration and validaton sets.Page 7, “Results”
- Table 7: Precision and recall across different part of speech (first POS: basis; second POS: derived word)Page 8, “Results”
- Table 7 shows precision and recall values for different part of speech combinations for the basis and derived words.Page 8, “Results”
- High precision and recall are achieved for NA derivations.Page 8, “Results”
See all papers in Proc. ACL 2013 that mention precision and recall.
See all papers in Proc. ACL that mention precision and recall.
Back to top.
rule-based
Appears in 8 sentences as: Rule-based (2) rule-based (6)
In DErivBase: Inducing and Evaluating a Derivational Morphology Resource for German
- This paper describes a rule-based framework for inducing derivational families (i.e., clusters of lemmas in derivational relationships) and its application to create a high-coverage German resource, DERIVBASE, mapping over 280k lemmas into more than 17k non-singleton clusters.Page 1, “Abstract”
- Instead, we employ a rule-based framework to define derivation rules that cover both suffixation and prefixation and describes stem changes.Page 2, “Introduction”
- Unsupervised approaches operate at the level of word-forms and have complementary strengths and weaknesses to rule-based approaches.Page 2, “Related Work”
- In this section, we describe our rule-based model of derivation, its operation to define derivational families, and the application of the model to German.Page 2, “Framework”
- As German is a morphologically complex language, we analyzed its derivation processes before implementing our rule-based model.Page 2, “Framework”
- 3.2 A Rule-based Derivation ModelPage 3, “Framework”
- Rule-based frameworks offer convenient representations for derivational morphology because they can take advantage of linguistic knowledge about derivation, have interpretable representations, and can be fine-tuned for high precision.Page 3, “Framework”
- In this paper, we present DERIVBASE, a derivational resource for German based on a rule-based framework.Page 9, “Conclusion and Future Work”
See all papers in Proc. ACL 2013 that mention rule-based.
See all papers in Proc. ACL that mention rule-based.
Back to top.
development set
Appears in 5 sentences as: development set (5)
In DErivBase: Inducing and Evaluating a Derivational Morphology Resource for German
- To this end, we constructed a development set comprised of a sample of 1,000 derivational families induced using our rules.Page 4, “Building the Resource”
- We also estimated the reliability of derivational rules by analyzing the accuracy of each rule on the development set .Page 5, “Building the Resource”
- We have considered a number of string distance measures and tested them on the development set (cf.Page 6, “Evaluation”
- This is based on preliminary experiments on the development set (cf.Page 6, “Evaluation”
- Lemmas included in the development set (Section 4.1) were excluded from sampling.Page 6, “Evaluation”
See all papers in Proc. ACL 2013 that mention development set.
See all papers in Proc. ACL that mention development set.
Back to top.
F1 score
Appears in 3 sentences as: F1 score (3)
In DErivBase: Inducing and Evaluating a Derivational Morphology Resource for German
- For the final evaluation, we optimized the number of clusters based on F1 score on calibration and validation sets (cf.Page 6, “Evaluation”
- We omit the F1 score because its use for precision and recall estimates from different samples is unclear.Page 7, “Results”
- Note that for these methods, precision and recall can be traded off against each other by varying the number of clusters; we chose the number of clusters by optimizing the F1 score on the calibration and validaton sets.Page 7, “Results”
See all papers in Proc. ACL 2013 that mention F1 score.
See all papers in Proc. ACL that mention F1 score.
Back to top.
