Please use this identifier to cite or link to this item:
https://ptsldigital.ukm.my/jspui/handle/123456789/513425Full metadata record
| DC Field | Value | Language |
|---|---|---|
| dc.contributor.advisor | Elankovan A. Sundararajan, Assoc. Prof. Dr. | - |
| dc.contributor.author | Mustafa Tareq Abd (P89127) | - |
| dc.date.accessioned | 2023-10-16T04:36:32Z | - |
| dc.date.available | 2023-10-16T04:36:32Z | - |
| dc.date.issued | 2021-04-05 | - |
| dc.identifier.other | ukmvital:130612 | - |
| dc.identifier.uri | https://ptsldigital.ukm.my/jspui/handle/123456789/513425 | - |
| dc.description | Density-based methods have appeared as a valuable approach for the clustering of evolving data streams. It has the potential to discover clusters of arbitrary shapes and to handle noise. Although several density-based algorithms have recently been developed for the clustering of evolving data streams, these algorithms are not without their issues. The first issue is the computational and memory efficiency required for clustering evolving data streams. From a computational perspective, grid-free evolving clustering leads to over granularity, which implies increased calling of the distance function and additional computation cost. From a memory aspect, grid-free clustering implies the allocation of a significant amount of memory. The second issue is the quality of the clustering is dramatically reduced when clustering high evolving nature of the data. While the third issue is an exponential increase in the processing time of traditional clustering algorithms for evolving data streams. In this study, new online clustering algorithms are developed to address these issues. The first proposed algorithm is called 'Clustering of Evolving Data streams via a density Grid-based Method' (CEDGM) and it consists of two fundamental steps such as, generate Core Micro-Clusters (CMC), and integration of any overlapping CMCs into macro clusters. The second proposed algorithm is called 'Clustering of Evolving Data stream via Grid-based Method with false Merging prevention' (CEDGM-Merge). It introduces two important parameters such as merge time interval" and "minimum number of links" to merge clusters that have a number of links above a certain level "minimum number of links" with the time that has passed on each one more than "merge time Interval". The third proposed algorithm is called 'Clustering of Evolving Data stream via Grid-based Method with Genetic Algorithm' (CEDGM-GA) to select suitable parameters successfully. This algorithm introduces the genetic algorithm (GA) to optimize the grid granularity and the minimum number of links parameters to the optimal value. The last proposed algorithm is called 'Parallel Clustering of Evolving Data stream via density Grid-based Method (P-CEDGM). It is a new algorithm for discovering clusters of evolving data streams implemented on Multi-Core CPU. In addition | - |
| 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 | Data streams | - |
| dc.subject | Grid-based method | - |
| dc.subject | Algorithms | - |
| dc.title | Online clustering of evolving data streams using density grid-based method in the internet of things | - |
| dc.type | Theses | - |
| dc.format.pages | 279 | - |
| dc.identifier.barcode | 005869(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_130612+Source01+Source010.PDF Restricted Access | 1.01 MB | Adobe PDF |  View/Open |
Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.