kReverse NN Search in Arbitrary Dimensionality

doi:10.1016/B978-012088469-8.50066-8

Proceedings 2004 VLDB Conference

2004, Pages 744-755

https://doi.org/10.1016/B978-012088469-8.50066-8 Get rights and content

Publisher Summary

This chapter focuses on conventional reverse Nearest Neighbor (RNN) queries. It also deals with reverse k-Nearest Neighbor (RkNN) queries. The existing algorithms for RNN search are applicable only in restricted scenarios. The chapter develops the first general methodology for retrieval of an arbitrary number of reverse nearest neighbors in multiple dimensions. This solution is better than the previous approaches in terms of efficiency and scalability. It illustrates several problem characteristics that permit the development of efficient algorithms and analyzes the performance of the proposed techniques with respect to existing methods. It introduces a two-step framework that retrieves a set of candidate RNNs and then removes the false misses. These two steps are independent and the filtering algorithms of one technique can be combined with the refinement mechanisms of the other.

References (0)

Cited by (260)

Predicting travel time within catchment area using Time Travel Voronoi Diagram (TTVD) and crowdsource map features
2022, Information Processing and Management
A catchment is a geographical area from which a business, service or organisation attracts its customers. A catchment area is a common way to ensure equal access to services such as hospitals, schools, libraries, ambulances, fire brigades, and shopping centres. Users will usually go to the service provider which is closest to their location instead of going further afield. In a time-sensitive environment where travelling time is limited, an incorrect decision might lead to serious consequences. In ambulance management, an incorrect dispatch that causes a unit late to arrive may lead to life and death situation. In this paper, we propose a Computational Geometry-based approach in determining catchment area, named Time Travel Voronoi Diagram (TTVD), not only by calculating the geographical location as used in most earlier work, but also through predicting the time travel to destination. This method can be used as a predictive analytic tool to support emergency dispatching, such as ambulance services. We utilise road features and the associated speed restrictions from crowdsource map platform in our prediction. Our simulation shows that a realistic catchment can be predicted using time-based distance with road features.
Multi-label sampling based on local label imbalance
2022, Pattern Recognition
Class imbalance is an inherent characteristic of multi-label data that hinders most multi-label learning methods. One efficient and flexible strategy to deal with this problem is to employ sampling techniques before training a multi-label learning model. Although existing multi-label sampling approaches alleviate the global imbalance of multi-label datasets, it is actually the imbalance level within the local neighbourhood of minority class examples that plays a key role in performance degradation. To address this issue, we propose a novel measure to assess the local label imbalance of multi-label datasets, as well as two multi-label sampling approaches, namely Multi-Label Synthetic Oversampling based on Local label imbalance (MLSOL) and Multi-Label Undersampling based on Local label imbalance (MLUL). By considering all informative labels, MLSOL creates more diverse and better labeled synthetic instances for difficult examples, while MLUL eliminates instances that are harmful to their local region. Experimental results on 13 multi-label datasets demonstrate the effectiveness of the proposed measure and sampling approaches for a variety of evaluation metrics, particularly in the case of an ensemble of classifiers trained on repeated samples of the original data.
Efficient processing of reverse nearest neighborhood queries in spatial databases
2020, Information Systems
This paper presents a novel query for spatial databases, called reverse nearest neighborhood (RNH) query, to discover the neighborhoods that find a query facility as their nearest facility among other facilities in the dataset. Unlike a reverse nearest neighbor (RNN) query, an RNH query emphasizes on group of users instead of an individual user. More specifically, given a set of user locations $U$ , a set of facility locations $F$ , a query location $q$ , a distance parameter $ρ$ and a positive integer $k$ , an RNH query returns all $ρ$ -radius circles $C$ enclosing at least $k$ users $u \in U$ , called neighborhoods (NH) such that the distance between $q$ and $C$ is less than the distance between $C$ and any other facility $f \in F$ . The RNH queries might have many practical applications including on demand facility placement and smart urban planning. We present an efficient approach for processing RNH queries on location data using R-tree based data indexing. In our approach, first we retrieve candidate RNH users by an efficient bound, prune and refine technique. Then, we incrementally discover RNHs of a query facility from these candidate RNH users. We also present the variants of RNH queries in spatial databases and propose solutions for them. We validate our approach by conducting extensive experiments with real datasets.
SAH: Shifting-Aware Asymmetric Hashing for Reverse k Maximum Inner Product Search
2023, Proceedings of the 37th AAAI Conference on Artificial Intelligence, AAAI 2023
Range-constrained probabilistic mutual furthest neighbor queries in uncertain databases
2023, Knowledge and Information Systems
Reverse spatial top-k keyword queries
2023, VLDB Journal

View all citing articles on Scopus

View full text

- kReverse NN Search in Arbitrary Dimensionality

Publisher Summary