Article Structure
Abstract
A process that attempts to solve abbreviation ambiguity is presented.
Introduction
An abbreviation is a letter or sequence of letters, which is a shortened form of a word or a sequence of words, which is called the sense of the abbreviation.
Abbreviation Disambiguation
The one sense per collocation hypothesis was introduced by Yarowsky (1993).
Abbreviation Disambiguation Features
Eighteen different features of any abbreviation instance were defined.
Implementing the OS Hypothesis
As mentioned above, the basic assumption of the OS hypothesis is that there exists at least one solvable abbreviation in the discourse and that the sense of that abbreviation is the same for all the instances of this abbreviation in the discourse.
Experiments
The examined dataset includes Jewish Law Documents written by two Jewish scholars: Rabbi Y. M. HaCohen (1995) and Rabbi O. Yosef (1977; 1986).
Conclusions, Summary and Future Work
This is the first ML system for disambiguation of abbreviations in Hebrew.
Topics
SVM
- An accuracy of 96.09% has been achieved by SVM .Page 1, “Abstract”
- : Maximum Entropy, SVM and C50.Page 2, “Abbreviation Disambiguation”
- Several well-known supervised ML methods have been selected: artificial neural networks (ANN), Nai've Bayes (NB), Support Vector Machines ( SVM ) and J48 (Witten and Frank, 1999) an improved variant of the C4.5 decision tree induction.Page 3, “Experiments”
- Table 2 shows that SVM achieved the best result with 96.09% accuracy.Page 3, “Experiments”
- ants ML Method ANN NB SVM J48Page 4, “Experiments”
- The comparison of the SVM results to the results of previous (Section 2) shows that our system achieves relatively high accuracy.Page 4, “Experiments”
See all papers in Proc. ACL 2008 that mention SVM.
See all papers in Proc. ACL that mention SVM.
Back to top.
best result
- Specifically, the AlWC_osWC feature variant achieves the best result with 87.75% accuracy.Page 3, “Experiments”
- Table 2 shows that SVM achieved the best result with 96.09% accuracy.Page 3, “Experiments”
- the achievements of four different standard ML methods, to the goal of achieving the best results , as opposed to the other systems that mainly focused on one ML method, each.Page 4, “Conclusions, Summary and Future Work”
See all papers in Proc. ACL 2008 that mention best result.
See all papers in Proc. ACL that mention best result.
Back to top.
natural languages
- The proposed system, preserves its portability between languages and domains because it does not use any natural language processing (NLP) subsystem (e.g.Page 1, “Introduction”
- This hypothesis states that natural languages tend to use consistent spoken and written styles.Page 1, “Abbreviation Disambiguation”
- hypothesis assumes that in natural languages , there is a tendency for an author to be consistent in the same discourse or article.Page 2, “Abbreviation Disambiguation”
See all papers in Proc. ACL 2008 that mention natural languages.
See all papers in Proc. ACL that mention natural languages.
Back to top.
