Please use this identifier to cite or link to this item:
https://ptsldigital.ukm.my/jspui/handle/123456789/513382Full metadata record
| DC Field | Value | Language |
|---|---|---|
| dc.contributor.advisor | Azuraliza Abu Bakar, Prof. Dr. | - |
| dc.contributor.author | Amin Mahmoudi (P85882) | - |
| dc.date.accessioned | 2023-10-16T04:36:05Z | - |
| dc.date.available | 2023-10-16T04:36:05Z | - |
| dc.date.issued | 2019-01-03 | - |
| dc.identifier.other | ukmvital:122013 | - |
| dc.identifier.uri | https://ptsldigital.ukm.my/jspui/handle/123456789/513382 | - |
| dc.description | Online social networks (OSNs) allow the detection of specific communities of users. This has led to the development of community detection (CD) algorithms. However, these algorithms are unable to define the time intervals needed to detect communities in time-varying OSNs; they are edge- and modularity-based so they only consider the number of connections between users, not the user attributes; and their computational complexity is high. OSN are dynamic and the key players are humans whose geo-location, density of interactions and user weight change over time and influence the formation of user communities. Therefore, this study aims to propose a new method to compute the time interval; a method to compute the user weight and a probability function of friendship based on geo-location; and a new CD algorithm based on the user attributes and time interval. Statistical functions are used to create a simulated model which is tested on six different datasets to identify the time interval. The user weight is computed by a simple exponential smoothing method which is tested on the Universiti Kebangsaan Malaysia (UKM) dataset and the result is compared with three existing methods using the pairwise F measure. Three large-scale datasets are analysed to determine the relation between geo-location and OSN ties. The proposed CD algorithm which is named Recently Largest Interaction (RLI) deploy gravitational search method and is tested on the Travian and UKM datasets and compared with the Dynamic Structural Clustering Algorithm for Network (DSCAN), edge-betweenness, label propagation, Walktrap, Infomap, leading eigenvector, and fast greedy algorithms using normalized mutual information, and adjusted rand index measures. The results show that the proposed approach is able to identify time interval accurately and subsequently the user weight computation method outperforms the three existing methods. In addition, estimating the probability of friendship based on the users attributes outperforms the scenario in which only geo-distance is considered, and above all, RLI algorithm can detect communities more accurately than existing algorithms and improve time and space complexity.,Ph.D. | - |
| dc.language.iso | eng | - |
| dc.publisher | UKM, Bangi | - |
| dc.relation | Faculty of Information Science and Technology / Fakulti Teknologi dan Sains Maklumat | - |
| dc.rights | UKM | - |
| dc.subject | Universiti Kebangsaan Malaysia -- Dissertations | - |
| dc.subject | Dissertations, Academic -- Malaysia | - |
| dc.subject | Online social networks | - |
| dc.subject | Social groups | - |
| dc.subject | Algorithms | - |
| dc.title | Temporal user attribute-base approach to detect communities in online social networks | - |
| dc.type | Theses | - |
| dc.format.pages | 228 | - |
| dc.identifier.callno | HM742.M334 2019 3 tesis | - |
| dc.identifier.barcode | 005536(2021)(PL2) | - |
| Appears in Collections: | Faculty of Information Science and Technology / Fakulti Teknologi dan Sains Maklumat | |
Files in This Item:
| File | Description | Size | Format | |
|---|---|---|---|---|
| ukmvital_122013+SOURCE1+SOURCE1.0.PDF Restricted Access | 3.12 MB | Adobe PDF |  View/Open |
Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.