عرض بسيط للتسجيلة

المؤلفBarham R.
المؤلفSharieh A.
المؤلفSleit A.
تاريخ الإتاحة2020-04-01T06:54:48Z
تاريخ النشر2019
اسم المنشورEvolutionary Intelligence
المصدرScopus
الرقم المعياري الدولي للكتاب18645909
معرّف المصادر الموحدhttp://dx.doi.org/10.1007/s12065-019-00257-y
معرّف المصادر الموحدhttp://hdl.handle.net/10576/13637
الملخصProviding a solution for the link prediction problem attracts several computer science fields and becomes a popular challenge in researches. This challenge is presented by introducing several approaches keen to provide the most precise prediction quality within a short period of time. The difficulty of the link prediction problem comes from the sparse nature of most complex networks such as social networks. This paper presents a parallel metaheuristic framework which is based on moth-flame optimization (MFO), clustering and pre-processed datasets to solve the link prediction problem. This framework is implemented and tested on a high-performance computing cluster and carried out on large and complex networks from different fields such as social, citation, biological, and information and publication networks. This framework is called Parallel MFO for Link Prediction (PMFO-LP). PMFO-LP is composed of data preprocessing stage and prediction stage. Dataset division with stratified sampling, feature extraction, data under-sampling, and feature selection are performed in the data preprocessing stage. In the prediction stage, the MFO based on clustering is used as the prediction optimizer. The PMFO-LP provides a solution to the link prediction problem with more accurate prediction results within a reasonable amount of time. Experimental results show that PMFO-LP algorithm outperforms other well-regarded algorithms in terms of error rate, the area under curve and speedup. Note that the source code of the PMFO-LP algorithm is available at https://github.com/RehamBarham/PMFO_MPI.cpp. - 2019, Springer-Verlag GmbH Germany, part of Springer Nature.
راعي المشروعThe authors would like to express their deep gratitude to IMAN1 Authority and the University of Jordan for their support in using their facilities.
اللغةen
الناشرSpringer Verlag
الموضوعComplex networks
Data clustering
Feature extraction
Link prediction problem
Moth-flame optimization
Parallel metaheuristic framework
العنوانMulti-moth flame optimization for solving the link prediction problem in complex networks
النوعArticle
الصفحات563-591
رقم العدد4
رقم المجلد12


الملفات في هذه التسجيلة

الملفاتالحجمالصيغةالعرض

لا توجد ملفات لها صلة بهذه التسجيلة.

هذه التسجيلة تظهر في المجموعات التالية

عرض بسيط للتسجيلة