應用於子維度天際線查詢的全新G-tree

A Novel G-tree for Subspace Skyline Query

天際線搜尋演算法在近期的資料庫研究領域中越趨重要。給定一組在多維度資料庫中的資料集，天際線搜尋會回傳那些不被其他點支配的資料點。在實務上，需要採用天際線搜尋的資料庫，通常都會提供多組候選維度，然而使用者只會對少部分有興趣。因此，通常都會根據多種維度的子集合進行查詢，而這類型的查詢稱之為子維度天際線查詢。使用傳統的天際線演算法來直接處理這種查詢是非常沒有效率的。有許多額外的演算法與架構被用來改善搜尋效率；然而，這些修改都會增加計算成本或是必須增加資料儲存空間。本篇論文提出一個基於Gaussian Function的全新索引模型，用來增強子維度天際線查詢的效率。實驗模擬結果展現了新的索引樹在子維度中找出天際線點的高效性。

The skyline search algorithm has recently emerged as an important technique in database research. Given a set of data points in a multidimensional database, such queries return points that are not 'dominated' by any other point. In practice, databases that require a skyline query usually provide numerous candidate dimensions, of which users are interested in only a few. As a result, queries are issued regarding various subsets of the dimensions and such queries are called subspace skyline queries. Using the conventional skyline algorithm to process these queries directly can be extremely ineffective. Additional algorithms and architectures have been added to improve search efficiency; however, such modifications can increase computational costs or necessitate an increase in data storage capacity. This paper proposes a novel index model based on a Gaussian function to enhance the performance of subspace skyline queries. Simulation results demonstrate the efficacy efficacy of the proposed tree in locating skyline points within a subspace.