Full Citation

Title: Bound-and-Filter Framework for Aggregate Reverse Rank Queries

Citation Type: Book, Section

Publication Year: 2018

ISBN:

ISSN:

DOI: 10.1007/978-3-662-58384-5_1

NSFID:

PMCID:

PMID:

Abstract: Finding top-rank products based on a given user’s preference is a user-view rank model that helps users to find their desired products. Recently, another query processing problem named reverse rank query has attracted significant research interest. The reverse rank query is a manufacturer-view model and can find users based on a given product. It can help to target potential users or find the placement for a specific product in marketing analysis. Unfortunately, previous reverse rank queries only consider one product, and they cannot identify the users for product bundling, which is known as a common sales strategy. To address the limitation, we propose a new query named aggregate reverse rank query to find matching users for a set of products. Three different aggregate rank functions (SUM, MIN, MAX) are proposed to evaluate a given product bundling in a variety of ways and target different users. To resolve these queries more efficiently, we propose a novel and sophisticated bound-and-filter framework. In the bound phase, two points are found to bound the query set for excluding candidates outside the bounds. In the filter phase, two tree-based methods are implemented with the bounds; they are the tree pruning method (TPM) and the double-tree method (DTM). The theoretical analysis and experimental results demonstrate the efficacy of the proposed methods.

Url: http://link.springer.com/10.1007/978-3-662-58384-5_1

User Submitted?: No

Authors: Dong, Yuyang; Chen, Hanxiong; Furuse, Kazutaka; Kitagawa, Hiroyuki

Editors:

Pages: 1-26

Volume Title: Transactions on Large-Scale Data- and Knowledge-Centered Systems XXXVIII

Publisher: Springer, Berlin, Heidelberg

Publisher Location: Heidelberg

Volume:

Edition:

Data Collections: IPUMS USA

Topics: Other

Countries: