Selectivity Estimation for Search Predicates over Set Valued Attributes

Коротков, А.С.; Kudryavtsev, Konstantin

doi:10.14257/ijdta.2016.9.10.24

Cited by 2 publications

(2 citation statements)

References 9 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Cannot well capture the set element dependency. Postgresql treats each element as a seperate attribute (column) and introduces a basic model and an extended model [GTK01] for selectivity estimation of set containment predicates [KK16,arr23]. The basic model does not take into account any element dependencies.…”

Section: Selectivity Estimation Involving Set-valued Attributesmentioning

confidence: 99%

“…They also only consider one set-valued attribute, so they cannot capture the attribute correlations. PostgreSQL treats each element as an attribute and introduce independence model or probabilistic model for set containment search predicates [KK16]. This method only captures the dependency of high-frequency elements and handles the queries on multiple attributes with attribute value independence (AVI) assumption.…”

mentioning

confidence: 99%

See 1 more Smart Citation

Enhancing the performance of selectivity estimation with machine learning techniques

Meng

View full text Add to dashboard Cite

I would like to express my deepest gratitude to my supervisor Prof. Gao Cong for providing patient guidance and unwavering support throughout my Ph.D journey. Over the past four years, his expertise, full of enthusiasm, carefulness and dedication always encouraged me on the road of academic research. He shaped the direction of my research and taught me a lot in the area of database system. I would also like to thank Prof. Xin Cao, Prof. Shuai Ma and Dr. Rong Zhu for their kind suggestions and valuable feedback to my research.I would also like to thank my colleagues and friends in our group, including Dr.

show abstract

Section: Selectivity Estimation Involving Set-valued Attributesmentioning

confidence: 99%

mentioning

confidence: 99%

Enhancing the performance of selectivity estimation with machine learning techniques

Meng

View full text Add to dashboard Cite

show abstract

Selectivity Estimation for Queries Containing Predicates over Set-Valued Attributes

Meng,

Cao,

Cong

2023

Proc. ACM Manag. Data

View full text Add to dashboard Cite

Selectivity estimation aims to estimate the size of query results size accurately and efficiently. Despite being a well-researched area for decades, most existing estimators are designed to handle comparison predicates over numeric and categorical data. Nevertheless, Set-valued data are ubiquitous in many applications such as information retrieval and recommendation systems. However, these estimators may not be effective for handling predicates over set-valued data. In this work, we presents novel techniques for selectivity estimation on queries involving predicates over set-valued attributes. We first propose the set-valued column factorization problem, whereby each each set-valued column is converted to multiple numeric subcolumns, and set containment predicates are converted to numeric comparison predicates. This enables us to leverage any existing estimator to perform selectivity estimation. We then develop two methods for column factorization and query conversion, namely ST and STH. We integrate ST and STH into three estimators, Postgres, Neurocard, and DeepDB. We then conduct a comprehensive empirical analysis by comparing our approach against three baselines across three different datasets. The experimental results demonstrate that our methods exhibit superior estimation accuracy while maintaining high efficiency compared to the baseline techniques.

show abstract

Selectivity Estimation for Search Predicates over Set Valued Attributes

Abstract: Abstract

Cited by 2 publications

References 9 publications

Enhancing the performance of selectivity estimation with machine learning techniques

Enhancing the performance of selectivity estimation with machine learning techniques

Selectivity Estimation for Queries Containing Predicates over Set-Valued Attributes

Contact Info

Product

Resources

About