Efficient Continuous kNN Join over Dynamic High-dimensional Data

Research Square (Research Square), Journal Year: 2023, Volume and Issue: unknown

Published: Aug. 24, 2023

Given a user dataset U and an object I, kNN join query in high-dimensional space returns the k nearest neighbors of each from I. The is basic necessary operation many applications, such as databases, data mining, computer vision, multi-media, machine learning, recommendation systems, more. In real world, datasets frequently update dynamically objects are added or removed. this paper, we propose novel methods continuous over dynamic highdimensional data. We firstly HDR+ Tree, which supports more efficient insertion, deletion, batch update. Further observed that existing rely on globally correlated for effective dimensionality reduction, then HDR Forest. It clusters constructs multiple Trees to capture local correlations among As result, our Forest able process non-globally efficiently. Two optimisations applied proposed Forest, including precomputation PCA states items pruning-based recomputation during item deletion. For completeness work, also present proof computing distances reduced dimensions Tree. Extensive experiments realworld show outperform baseline algorithms naive RkNN

Language: Английский

---

Efficient Continuous kNN Join over Dynamic High-dimensional Data

Research Square (Research Square), Journal Year: 2023, Volume and Issue: unknown

Published: Aug. 25, 2023

Given a user dataset U and an object I, kNN join query in high-dimensional space returns the k nearest neighbors of each from I. The is basic necessary operation many applications, such as databases, data mining, computer vision, multi-media, machine learning, recommendation systems, more. In real world, datasets frequently update dynamically objects are added or removed. this paper, we propose novel methods continuous over dynamic highdimensional data. We firstly HDR+ Tree, which supports more efficient insertion, deletion, batch update. Further observed that existing rely on globally correlated for effective dimensionality reduction, then HDR Forest. It clusters constructs multiple Trees to capture local correlations among As result, our Forest able process non-globally efficiently. Two optimisations applied proposed Forest, including precomputation PCA states items pruning-based recomputation during item deletion. For completeness work, also present proof computing distances reduced dimensions Tree. Extensive experiments realworld show outperform baseline algorithms naive RkNN

Language: Английский

---

Efficient Continuous kNN Join over Dynamic High-dimensional Data

Research Square (Research Square), Journal Year: 2023, Volume and Issue: unknown

Published: Sept. 18, 2023

Given a user dataset U and an object I, kNN join query in high-dimensional space returns the k nearest neighbors of each from I. The is basic necessary operation many applications, such as databases, data mining, computer vision, multi-media, machine learning, recommenda-tion systems, more. In real world, datasets frequently update dynamically objects are added or removed. this paper, we propose novel methods continuous over dynamic data. We firstly HDR+ Tree, which supports more efficient insertion, deletion, batch update. Further observed that existing rely on globally correlated for effec-tive dimensionality reduction, then HDR Forest. It clusters constructs multiple Trees to capture local correlations among As result, our Forest able process non-globally efficiently. Two optimisations applied proposed Forest, including precomputation PCA states items pruning-based recomputation during item deletion. For completeness work, also present proof computing distances reduced dimensions Tree. Extensive experiments real-world show outperform baseline algorithms naive RkNN

Language: Английский

---