Semantic Query Expansion for Arabic Information Retrieval

Mahgoub, Ashraf; Rashwan, Mohsen; Raafat, Hazem M.; Zahran, Mohamed; Fayek, Magda B.

doi:10.3115/v1/w14-3611

Cited by 27 publications

(18 citation statements)

References 3 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…In general, the studies that addressed the support of Arabic language on the SW can be divided into four categories [7]: 1) the development of Arabic ontologies [18,20,21], 2) Employing ontologies to improve Arabic named entities extraction [4,36], 3) Ontology based representation of Islamic knowledge [2,19,24] and 4) supporting cross-language information retrieval and search [17,33]. Although an increasing number of efforts have started to use ontologies to enhance information retrieval from Arabic data [29,31], the use of ontologies was almost limited to query expansion, and results were retrieved from unstructured data on the Web. Our work takes a different direction by addressing NL interfaces for querying ontologies and RDF stores.…”

Section: Support For Arabic Language On the Swmentioning

confidence: 99%

AR2SPARQL: An Arabic Natural Language Interface for the Semantic Web

AlAgha¹,

Abu-Taha²

2015

IJCA

View full text Add to dashboard Cite

With the growing interest in supporting the Arabic language on the Semantic Web (SW), there is an emerging need to enable Arab users to query ontologies and RDF stores without being challenged with the formal logic of the SW. In the domain of English language, several efforts provided Natural Language (NL) interfaces to enable ordinary users to query ontologies using NL queries. However, none of these efforts were designed to support the Arabic language which has different morphological and semantic structures.As a step towards supporting Arabic Question Answering (QA) on the SW, this work presents AR2SPARQL, a NL interface that takes questions expressed in Arabic and returns answers drawn from an ontology-based knowledge base. The core of AR2SPARQL is the approach we propose to translate Arabic questions into triples which are matched against RDF data to retrieve an answer. The system uses both linguistic and semantic features to resolve ambiguity when matching words to the ontology content. To overcome the limited support for Arabic Natural Language Processing (NLP), the system does not make intensive use of sophisticated linguistic methods. Instead, it relies more on the knowledge defined in the ontology and the grammar rules we define to capture the structures of Arabic questions and to construct an adequate RDF representations. AR2SPARQL has been tested with two different datasets and results have shown that it achieves a good retrieval performance in terms of precision and recall.

show abstract

Section: Support For Arabic Language On the Swmentioning

confidence: 99%

AR2SPARQL: An Arabic Natural Language Interface for the Semantic Web

AlAgha¹,

Abu-Taha²

2015

IJCA

View full text Add to dashboard Cite

show abstract

“…Wikipedia-based categories have been also exploited to improve the categorization of Arabic text [32]. Some works have also used Arabic Wikipedia as background knowledge to expand queries submitted to search engines or question answering systems [33]. The work in this paper adds to previous knowledge by extending the use of Arabic Wikipedia to include the entity-linking problem.…”

Section: Related Workmentioning

confidence: 79%

Automatic Linking of Short Arabic Texts to Wikipedia Articles

Fayad¹,

AlAgha²

2016

JSW

View full text Add to dashboard Cite

Abstract:Given the enormous amount of unstructured texts available on the Web, there has been an emerging need to increase discoverability of and accessibility to these texts. One of the proposed solutions is to annotate texts with information extracted from background knowledge. Wikipedia, the free encyclopedia, has been recently exploited as a background knowledge to annotate text with complementary information. Given any piece of text, the main challenge is how to determine the most relevant information from Wikipedia with the least effort and time. While Wikipedia-based annotation has mainly targeted the English and Latin versions of Wikipedia, little effort has been devoted to annotate Arabic text using the Arabic version of Wikipedia. In addition, the annotation of short text presents further challenges due to the inability to apply statistical or machine learning techniques that are commonly used with long text. This work proposes an approach for automatic linking of Arabic short texts to articles drawn from Wikipedia. It reports on the several challenges associated with the design and implementation of the linking approach including the processing of the Wikipedia's enormous content, the mapping of texts to Wikipedia articles, the problem of article disambiguation, and the time efficiency. The proposed approach was tested on a dataset of 100 short texts gathered from online Arabic articles. The annotations generated by the approach were compared with the annotations generated by two human subjects. The approach achieved 71.79% accuracy, 74.70% average precision, and 82.63 % average recall. A thorough analysis and discussion of the evaluation results are also presented to address the limitations, strengths as well as recommendations for future improvements.

show abstract

“…Here, we propose using the Arabic vectors as a semantic expansion technique because the Arabic vectors capture the semantic properties of the language such that semantically close terms are clustered in close proximity in the vector space. Mahgoub et al [ 23] proposed query semantic expansion techniques for Arabic information retrieval, the expansion techniques based on various language resources as Wikipedia, Google translate with WordNet, and other various Arabic linguistic resources. We compare our vector expansion technique with their techniques using TREC 2002 the cross-lingual (CLIR) track dataset [ 24], which contains 50 queries tested against 383,872 documents, we discarded any non-judged documents from our experiments before evaluation.…”

Section: Information Retrievalmentioning

confidence: 99%

“…In order to avoid bias in reordering the expansion list, the term being expanded is not included among the terms forming the query vector. For each query, we allow maximum 50 expansions for all of its terms, such that the number of expansions for each term is inversely proportional to its frequency, thus allowing less frequent terms to have more expansions [ 23]. Figure 1 and 2 compare between the impact of using the Arabic vectors as an expansion scheme versus traditional resources as Wikipedia, WordNet translations and other resources using Indiri [ 25].…”

Section: Information Retrievalmentioning

confidence: 99%

Word Representations in Vector Space and their Applications for Arabic

Zahran

Magooda

Mahgoub

et al. 2015

Lecture Notes in Computer Science

Self Cite

View full text Add to dashboard Cite

Abstract.A lot of work has been done to give the individual words of a certain language adequate representations in vector space so that these representations capture semantic and syntactic properties of the language. In this paper, we compare different techniques to build vectorized space representations for Arabic, and test these models via intrinsic and extrinsic evaluations. Intrinsic evaluation assesses the quality of models using benchmark semantic and syntactic dataset, while extrinsic evaluation assesses the quality of models by their impact on two Natural Language Processing applications: Information retrieval and Short Answer Grading. Finally, we map the Arabic vector space to the English counterpart using Cosine error regression neural network and show that it outperforms standard mean square error regression neural networks in this task.

show abstract

Semantic Query Expansion for Arabic Information Retrieval

Cited by 27 publications

References 3 publications

AR2SPARQL: An Arabic Natural Language Interface for the Semantic Web

AR2SPARQL: An Arabic Natural Language Interface for the Semantic Web

Automatic Linking of Short Arabic Texts to Wikipedia Articles

Word Representations in Vector Space and their Applications for Arabic

Contact Info

Product

Resources

About