Efficient k-Regret Query Algorithm with Restriction-free Bound for any Dimensionality

Xie, Min; Wong, Raymond Chi-Wing; Li, Jian; Long, Cheng; Lall, Ashwin

doi:10.1145/3183713.3196903

Cited by 47 publications

(50 citation statements)

References 28 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The convex hull has also been applied in multiple studies and fields. The algorithm SPHERE [22] used a convex hull for the k-regret query to attain the lower bound of the maximum regret ratio. Meanwhile, Peng et al constructed a convex hull to find candidate points for the k-regret query [23].…”

Section: Convex Hull Methodsmentioning

confidence: 99%

UB-H: an unbalanced-hierarchical layer binary-wise construction method for high-dimensional data

Ihm

Park

2021

Computing

View full text Add to dashboard Cite

Cloud computing, which is distributed, stored and managed, is drawing attention as data generation and storage volumes increase. In addition, research on green computing, which increases energy efficiency, is also widely studied. An index is constructed to retrieve huge dataset efficiently, and the layer-based indexing methods are widely used for efficient query processing. These methods construct a list of layers, so that only one layer is required for information retrieval instead of the entire dataset. The existing layer-based methods construct the layers using a convex hull algorithm. However, the execution time of this method is very high, especially in large, high-dimensional datasets. Furthermore, if the total number of layers increases, the query processing time also increases, resulting in efficient, but slow, query processing. In this paper, we propose an unbalanced-hierarchical layer method, which hierarchically divides the dimensions of input data to increase the total number of layers and reduce the index building time. We demonstrate that the proposed procedure significantly increases the total number of layers and reduces the index building time, compared to existing methods through the various experiments.

show abstract

Section: Convex Hull Methodsmentioning

confidence: 99%

UB-H: an unbalanced-hierarchical layer binary-wise construction method for high-dimensional data

Ihm

Park

2021

Computing

View full text Add to dashboard Cite

show abstract

“…al [15] proposed a space-partition based algorithm which returns a set of tuples whose minimum happiness ratio is lower bounded. This bound was improved in [8], [26] which is derived based on the well-known ǫkernel problem [1]. Greedy-based algorithms which construct the solutions iteratively were proposed in [9], [15], [16], [26].…”

Section: R Elated Workmentioning

confidence: 99%

“…This bound was improved in [8], [26] which is derived based on the well-known ǫkernel problem [1]. Greedy-based algorithms which construct the solutions iteratively were proposed in [9], [15], [16], [26]. Besides, the k-regret query can be solved in a set-cover manner [4] while user interactions were considered in [14], [24], [25].…”

Section: R Elated Workmentioning

confidence: 99%

“…Some k-regret query algorithms can be modified to answer an α-happiness query. One could set a proper k in the theoretically bounded algorithms [8], [15], [26] so that the lower bound on the minimum happiness ratio is at least α. Unfortunately, the value of k set in this approach can be extremely large.…”

Section: R Elated Workmentioning

confidence: 99%

“…For example, to guarantee 0.95 happy on a 3-dimensional dataset, the space-partition based algorithm in [15] has to return 1400 tuples, which is unacceptable in real scenarios. One could also modify the greedy-based algorithms [9], [15], [16], [26] so as to include tuples greedily until the minimum happiness level of users is satisfied (we consider this approach in our experiments). Similarly, solutions for the α-happiness query can also be extended to answer the kregret query: using an approach discussed in [3], [13], one can perform a binary search on different happiness ratios to obtain a solution set whose output size is at most k.…”

Section: R Elated Workmentioning

confidence: 99%

See 2 more Smart Citations