Building a Hierarchical Annotated Corpus of Urdu: The URDU.KON-TB Treebank

Abbas, Qaiser

doi:10.1007/978-3-642-28604-9_6

Cited by 11 publications

(28 citation statements)

References 12 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Unfortunately, the document of the Urdu 5000 most frequently used words contained only counts of words. So, the probability value for each word listed in the document was calculated using (1). A sample of unigram and probability values is given in Fig.…”

Section: A N-gram Acquisitionmentioning

confidence: 99%

“…As Urdu is an under-resourced language and no such precise or sufficient support available on today's machines for this language. This work presented here is a positive 1 http://www.autohotkey.com/community/viewtopic.php?f=2\&t=53630 2 http://www.clasohm.com/lmt/en/ 3 http://jsimlo.sk/notepad/ contribution for Urdu word prediction (UWP) in general and a helpful tool to boost up the typing needs of related handicapped persons.…”

Section: Introductionmentioning

confidence: 99%

“…In near future, the extension of this QASKU model towards the PST's approach will be possible because a treebank for the Urdu language is under construction by Abbas [1]. So, it was decided to move forward in consecutive steps.…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

A Stochastic Prediction Interface for Urdu

Abbas

2014

IJISA

Self Cite

View full text Add to dashboard Cite

Abstract-This work lays down a foundation for text prediction of an inflected and under-resourced language Urdu. The interface developed is not limited to a T9 (Text on 9 keys) application used in embedded devices, which can only predict a word after typing initial characters. It is capable of predicting a word like T9 and also a sequence of word after a word in a continuous manner for fast document typing. It is based on Ngram language model. This stochastic interface deals with three N-gram levels from unary to ternary independently. The unigram mode is being in use for applications like T9, while the bigram and tri-gram modes are being in use for sentence prediction. The measures include a percentage of keystrokes saved, keystrokes until completion and a percentage of time saved during the typing. Two different corpora are merged to build a sufficient amount of data. The test data is divided into a test and a held out data equally for an experimental purpose. This whole exercise enables the QASKU system outperforms the FastType with almost 15% more saved keystrokes.

show abstract

Section: A N-gram Acquisitionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

A Stochastic Prediction Interface for Urdu

Abbas

2014

IJISA

Self Cite

View full text Add to dashboard Cite

show abstract

“…A comparative study made is detailed in Section 5. The URDU.KON-TB treebank having phrase structure (PS) and the hyper dependency structure (HDS) annotation with rich encoded information (Abbas, 2012;Abbas, 2014) is used for the training of the Urdu parser discussed in Section 3. The treebank has a semi-semantic POS (SSP) tag set, a semi-semantic syntactic (SSS) tag set and a functional (F) tag set.…”

Section: Related Workmentioning

confidence: 99%

Exploiting Language Variants Via Grammar Parsing Having Morphologically Rich Information

Abbas

2014

Proceedings of the EMNLP'2014 Workshop on Language Technology for Closely Related Languages and Language Variants

Self Cite

View full text Add to dashboard Cite

In this paper, the development and evaluation of the Urdu parser is presented along with the comparison of existing resources for the language variants Urdu/Hindi. This parser was given a linguistically rich grammar extracted from a treebank. This context free grammar with sufficient encoded information is comparable with the state of the art parsing requirements for morphologically rich and closely related language variants Urdu/Hindi. The extended parsing model and the linguistically rich grammar together provide us promising parsing results for both the language variants. The parser gives 87% of f-score, which outperforms the multi-path shift-reduce parser for Urdu and a simple Hindi dependency parser with 4.8% and 22% increase in recall, respectively.

show abstract

“…Despite remarkable work on text categorization of English documents, the work on Urdu language text is still in infancy [32] though online Urdu text data is rapidly growing and necessitating the need to develop methods to organize and handle data. An important reason for this lack of interest is unavailability of publically accessible Urdu data collections [23,30].…”

Section: Introductionmentioning

confidence: 99%

Evaluation of Feature Selection Approaches for Urdu Text Categorization

Zia¹,

Abbas²,

Akhtar³

2015

IJISA

Self Cite

View full text Add to dashboard Cite

Abstract-Efficient feature selection is an important phase of designing an effective text categorization system. Various feature selection methods have been proposed for selecting dissimilar feature sets. It is often essential to evaluate that which method is more effective for a given task and what size of feature set is an effective model selection choice. Aim of this paper is to answer these questions for designing Urdu text categorization system. Five widely used feature selection methods were examined using six well-known classification algorithms: naive Bays (NB), k-nearest neighbor (KNN), support vector machines (SVM) with linear, polynomial and radial basis kernels and decision tree (i.e. J48). The study was conducted over two test collections: EMILLE collection and a naive collection. We have observed that three feature selection methods i.e. information gain, Chi statistics, and symmetrical uncertain, have performed uniformly in most of the cases if not all. Moreover, we have found that no single feature selection method is best for all classifiers. While gain ratio out-performed others for naive Bays and J48, information gain has shown top performance for KNN and SVM with polynomial and radial basis kernels. Overall, linear SVM with any of feature selection methods including information gain, Chi statistics or symmetric uncertain methods is turned-out to be first choice across other combinations of classifiers and feature selection methods on moderate size naive collection. On the other hand, naive Bays with any of feature selection method have shown its advantage for a small sized EMILLE corpus.

show abstract

Building a Hierarchical Annotated Corpus of Urdu: The URDU.KON-TB Treebank

Cited by 11 publications

References 12 publications

A Stochastic Prediction Interface for Urdu

A Stochastic Prediction Interface for Urdu

Exploiting Language Variants Via Grammar Parsing Having Morphologically Rich Information

Evaluation of Feature Selection Approaches for Urdu Text Categorization

Contact Info

Product

Resources

About