Evaluation of different information retrieval models and different indexing methods using different similarity measures on Arabic documents

مقدم أطروحة جامعية

Hanandeh, Isam Said S.

مشرف أطروحة جامعية

Kanan, Ghassan Jaddu

الجامعة

الأكاديمية العربية للعلوم المالية و المصرفية

الكلية

كلية نظم و تكنولوجيا المعلومات

القسم الأكاديمي

قسم نظم المعلومات الحاسوبية

دولة الجامعة

الأردن

الدرجة العلمية

دكتوراه

تاريخ الدرجة العلمية

2008

الملخص الإنجليزي

Information retrieval deals with representation, storage, organizational of and access to information items.

Information retrieval is a difficult problem for some factors. We have huge amount of data, the desire to keep related items together, documents have unstructured domain and data and the information is distributed and interlinked not forgetting the performance criteria. As the volume of data and documents has reached a limit which is statistically unsuitability, and as the search processes, whether in the Internet or libraries have become more complicated, it is necessary to improve the search process.

It is worthy to mention that studies and research executed in English languages in this field are abundant.

However, Arabic language efforts are still limited.

The present study is, Investigate how the automatic indexing work, we applied this work on Arabic language information retrieval systems.

The equations used are based on the statistical analysis of documents ; word occurrences in the document and in the remaining documents are the inputs to these equations. Through other equations search expressions are assigned weights.

Also the degree of similarity between the expressions is calculated.

Equations necessary for analyzing the documents, calculating weights have been programmed.

As result after applied.

In this research we need to know which automatic indexing technique significant in Arabic information retrieval system. We ran the system over 242 Arabic documents and 60 predefined Arabic natural language queries from the proceeding of the Saudi Arabian national computer conference.

This thesis present some aims : Design and build different Information retrieval models and investigate the applicability of these model in Arabic information retrieval, evaluation of different information retrieval indexing methods using different similarity measures on Arabic documents.

Finally Applying the work on Arabic information retrieval system using Arabic root and Arabic full word as index terms and applying the work in inverted file once in block oriented mechanism and the other in word oriented mechanism. Through implementing our work over the measures, we will see that the best results are always coming from searching by inverted file over the other file and the search by inverted file using root is the better result than search by full word and the search using block oriented mechanism is give better result than search by word oriented mechanism due to search process and measures.

التخصصات الرئيسية

تكنولوجيا المعلومات وعلم الحاسوب

الموضوعات

• استرجاع المعلومات

عدد الصفحات

104

قائمة المحتويات

• APA
• MLA
• AMA

لغة النص

الإنجليزية

نوع البيانات

رسائل جامعية

رقم السجل

BIM-306238