Continuous Top-k Dominating Query of Incomplete Data over Data Streams

Santoso, Bagus Jati; Permadi, Vynska Amalia; Ahmad, Tohari; Ijtihadie, Royyana Muslim; Sektiaji, Bayu

doi:10.1109/siet.2018.8693162

Cited by 4 publications

(1 citation statement)

References 6 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…A dynamic dominate model to get the largest number of uncertain objects is proposed in [14] from uncertain objects to reduce the search space to answer the queries by a pruning approach. The dynamic environment is also handled in [15] for incomplete data by event-based method to answer top-k dominating query. A parallel approach for processing TKD queries on incomplete data using MapReduce is proposed in [16].…”

Section: Related Workmentioning

confidence: 99%

Indexed Top-k Dominating Queries on Highly Incomplete Data

Fattah

Hasan

Tsuji

2021

Lecture Notes on Data Engineering and Communications Technologies

View full text Add to dashboard Cite

Top-k dominating (TKD) query returns the top-k data items in a dataset that dominate other objects. This is an important decision-making tool for any business organizations because it gives data analysts an insightful way to find dominating objects. It incorporates the benefits of skyline and top-k queries and is used in a variety of decision-making applications. Due to system malfunction, privacy protection, data loss, and other factors, incomplete data occurs in a wide range of actual applications. We design an algorithm for answering the top-k dominating queries applying the idea of data bucketing. Each of the buckets is implemented with a B+ tree. A key is generated for each of the incomplete data record and inserted into the B+ tree. Using the key, we reduce the computation cost for comparison applying the B+ index structure. Since the B+ tree is well accepted for efficient indexing for commercial databases, we get the facility of high retrieval performance. The proposed approach clearly outperforms in terms of query performance compared to many other existing approaches for incomplete data.

show abstract

Section: Related Workmentioning

confidence: 99%