Efficient skyline processing algorithm over dynamic and incomplete database
The notion of skyline processing is to discover the data items that are not dominated by any other data items. It is a well-known technique that is utilised to determine the best results that meet the user’s preferences. However, the rapid growth and frequent changes of data make the process of ide...
| Main Authors: | , , , , |
|---|---|
| Format: | Conference or Workshop Item |
| Language: | English English |
| Published: |
ACM
2018
|
| Subjects: | |
| Online Access: | http://irep.iium.edu.my/69868/ http://irep.iium.edu.my/69868/ http://irep.iium.edu.my/69868/ http://irep.iium.edu.my/69868/7/69868_Efficient%20skyline%20processing%20algorithm_complete.pdf http://irep.iium.edu.my/69868/8/69868_Efficient%20skyline%20processing%20algorithm_scopus.pdf |
| Summary: | The notion of skyline processing is to discover the data items that are not dominated by any other data items. It is a well-known technique that is utilised to determine the best results that meet the user’s preferences. However, the rapid growth and frequent changes of data make the process of identifying skyline points no longer a trivial task. Most of the existing skyline approaches assume that the database is complete and static. However, in real world scenario, this assumption is not valid especially in multidimensional databases in which some dimensions have missing values while they are dynamic due to the continual modifications made towards them. Blindly examining the whole database after changes are made to identify the skyline points is inappropriate as not all data items are affected by the changes. Hence, in this study we propose a skyline algorithm, DyIn-Skyline, which is capable of identifying skyline points over dynamic and incomplete databases, by exploiting only those data items that are affected by the changes. Several experiments have been conducted and the results show that our proposed algorithm outperforms the previous work by reducing the number of pairwise comparisons in the range of 50% to 73%. |
|---|