المصدر: | مجلة جامعة الملك سعود - علوم الحاسب والمعلومات |
---|---|
الناشر: | جامعة الملك سعود |
المؤلف الرئيسي: | Khan, Sharifullah (Author) |
مؤلفين آخرين: | Mustafa, Jibran (Co-Author) |
المجلد/العدد: | مج26, ع2 |
محكمة: | نعم |
الدولة: |
السعودية |
التاريخ الميلادي: |
2014
|
الصفحات: | 161 - 169 |
DOI: |
10.33948/0584-026-002-001 |
ISSN: |
1319-1578 |
رقم MD: | 973064 |
نوع المحتوى: | بحوث ومقالات |
اللغة: | الإنجليزية |
قواعد المعلومات: | science |
مواضيع: | |
كلمات المؤلف المفتاحية: |
Semantic Search | Thematic Similarity | Semantic Heterogeneity | RDF Triples | Information Retrieval
|
رابط المحتوى: |
المستخلص: |
Most existing semantic search systems expand search keywords using domain ontology to deal with semantic heterogeneity. They focus on matching the semantic similarity of individual keywords in a multiple-keywords query; however, they ignore the semantic relationships that exist among the keywords of the query themselves. The systems return less relevant answers for these types of queries. More relevant documents for a multiple-keywords query can be retrieved if the systems know the relationships that exist among multiple keywords in the query. The proposed search methodology matches patterns of keywords for capturing the context of keywords, and then the relevant documents are ranked according to their pattern relevance score. A prototype system has been implemented to validate the proposed search methodology. The system has been compared with existing systems for evaluation. The results demonstrate improvement in precision and recall of search. |
---|---|
ISSN: |
1319-1578 |