RecStore: An extensible and adaptive framework for online recommender queries inside the database engine

Justin J. Levandoski, Mohamed Sarwat, Mohamed Mokbel, Michael D. Ekstrand

Research output: Chapter in Book/Report/Conference proceedingConference contribution

5 Citations (Scopus)

Abstract

Most recommendation methods (e.g., collaborative filtering) consist of (1) a computationally intense offline phase that computes a recommender model based on users' opinions of items, and (2) an online phase consisting of SQL-based queries that use the model (generated offline) to derive user preferences and provide recommendations for interesting items. Current application usage trends require a completely online recommender process, meaning the recommender model must update in real time as new opinions enter the system. To tackle this problem, we propose RecStore, a DBMS storage engine module capable of efficient online model maintenance. Externally, models managed by RecStore behave as relational tables, thus existing SQL-based recommendation queries remain unchanged while gaining online model support. RecStore maintains internal statistics and data structures aimed at providing efficient incremental updates to the recommender model, while employing an adaptive strategy for internal maintenance and load shedding to realize a balance between efficiency in updates or query processing based on system workloads. RecStore is also extensible, supporting a declarative syntax for defining recommender models. The efficacy of RecStore is demonstrated by providing the implementation details of three state-of-the-art collaborative filtering models. We provide an extensive experimental evaluation of a prototype of RecStore, built inside the storage engine of PostgreSQL, using a real-life recommender system workload.

Original languageEnglish
Title of host publicationAdvances in Database Technology - EDBT 2012
Subtitle of host publication15th International Conference on Extending Database Technology, Proceedings
Pages86-96
Number of pages11
DOIs
Publication statusPublished - 10 Jul 2012
Externally publishedYes
Event15th International Conference on Extending Database Technology, EDBT 2012 - Berlin, Germany
Duration: 27 Mar 201230 Mar 2012

Other

Other15th International Conference on Extending Database Technology, EDBT 2012
CountryGermany
CityBerlin
Period27/3/1230/3/12

Fingerprint

Engines
Collaborative filtering
Query processing
Recommender systems
Data structures
Statistics

ASJC Scopus subject areas

  • Software
  • Human-Computer Interaction
  • Computer Vision and Pattern Recognition
  • Computer Networks and Communications

Cite this

Levandoski, J. J., Sarwat, M., Mokbel, M., & Ekstrand, M. D. (2012). RecStore: An extensible and adaptive framework for online recommender queries inside the database engine. In Advances in Database Technology - EDBT 2012: 15th International Conference on Extending Database Technology, Proceedings (pp. 86-96) https://doi.org/10.1145/2247596.2247608

RecStore : An extensible and adaptive framework for online recommender queries inside the database engine. / Levandoski, Justin J.; Sarwat, Mohamed; Mokbel, Mohamed; Ekstrand, Michael D.

Advances in Database Technology - EDBT 2012: 15th International Conference on Extending Database Technology, Proceedings. 2012. p. 86-96.

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Levandoski, JJ, Sarwat, M, Mokbel, M & Ekstrand, MD 2012, RecStore: An extensible and adaptive framework for online recommender queries inside the database engine. in Advances in Database Technology - EDBT 2012: 15th International Conference on Extending Database Technology, Proceedings. pp. 86-96, 15th International Conference on Extending Database Technology, EDBT 2012, Berlin, Germany, 27/3/12. https://doi.org/10.1145/2247596.2247608
Levandoski JJ, Sarwat M, Mokbel M, Ekstrand MD. RecStore: An extensible and adaptive framework for online recommender queries inside the database engine. In Advances in Database Technology - EDBT 2012: 15th International Conference on Extending Database Technology, Proceedings. 2012. p. 86-96 https://doi.org/10.1145/2247596.2247608
Levandoski, Justin J. ; Sarwat, Mohamed ; Mokbel, Mohamed ; Ekstrand, Michael D. / RecStore : An extensible and adaptive framework for online recommender queries inside the database engine. Advances in Database Technology - EDBT 2012: 15th International Conference on Extending Database Technology, Proceedings. 2012. pp. 86-96
@inproceedings{0595ca117edd473b86f345fdffb9fb63,
title = "RecStore: An extensible and adaptive framework for online recommender queries inside the database engine",
abstract = "Most recommendation methods (e.g., collaborative filtering) consist of (1) a computationally intense offline phase that computes a recommender model based on users' opinions of items, and (2) an online phase consisting of SQL-based queries that use the model (generated offline) to derive user preferences and provide recommendations for interesting items. Current application usage trends require a completely online recommender process, meaning the recommender model must update in real time as new opinions enter the system. To tackle this problem, we propose RecStore, a DBMS storage engine module capable of efficient online model maintenance. Externally, models managed by RecStore behave as relational tables, thus existing SQL-based recommendation queries remain unchanged while gaining online model support. RecStore maintains internal statistics and data structures aimed at providing efficient incremental updates to the recommender model, while employing an adaptive strategy for internal maintenance and load shedding to realize a balance between efficiency in updates or query processing based on system workloads. RecStore is also extensible, supporting a declarative syntax for defining recommender models. The efficacy of RecStore is demonstrated by providing the implementation details of three state-of-the-art collaborative filtering models. We provide an extensive experimental evaluation of a prototype of RecStore, built inside the storage engine of PostgreSQL, using a real-life recommender system workload.",
author = "Levandoski, {Justin J.} and Mohamed Sarwat and Mohamed Mokbel and Ekstrand, {Michael D.}",
year = "2012",
month = "7",
day = "10",
doi = "10.1145/2247596.2247608",
language = "English",
isbn = "9781450307901",
pages = "86--96",
booktitle = "Advances in Database Technology - EDBT 2012",

}

TY - GEN

T1 - RecStore

T2 - An extensible and adaptive framework for online recommender queries inside the database engine

AU - Levandoski, Justin J.

AU - Sarwat, Mohamed

AU - Mokbel, Mohamed

AU - Ekstrand, Michael D.

PY - 2012/7/10

Y1 - 2012/7/10

N2 - Most recommendation methods (e.g., collaborative filtering) consist of (1) a computationally intense offline phase that computes a recommender model based on users' opinions of items, and (2) an online phase consisting of SQL-based queries that use the model (generated offline) to derive user preferences and provide recommendations for interesting items. Current application usage trends require a completely online recommender process, meaning the recommender model must update in real time as new opinions enter the system. To tackle this problem, we propose RecStore, a DBMS storage engine module capable of efficient online model maintenance. Externally, models managed by RecStore behave as relational tables, thus existing SQL-based recommendation queries remain unchanged while gaining online model support. RecStore maintains internal statistics and data structures aimed at providing efficient incremental updates to the recommender model, while employing an adaptive strategy for internal maintenance and load shedding to realize a balance between efficiency in updates or query processing based on system workloads. RecStore is also extensible, supporting a declarative syntax for defining recommender models. The efficacy of RecStore is demonstrated by providing the implementation details of three state-of-the-art collaborative filtering models. We provide an extensive experimental evaluation of a prototype of RecStore, built inside the storage engine of PostgreSQL, using a real-life recommender system workload.

AB - Most recommendation methods (e.g., collaborative filtering) consist of (1) a computationally intense offline phase that computes a recommender model based on users' opinions of items, and (2) an online phase consisting of SQL-based queries that use the model (generated offline) to derive user preferences and provide recommendations for interesting items. Current application usage trends require a completely online recommender process, meaning the recommender model must update in real time as new opinions enter the system. To tackle this problem, we propose RecStore, a DBMS storage engine module capable of efficient online model maintenance. Externally, models managed by RecStore behave as relational tables, thus existing SQL-based recommendation queries remain unchanged while gaining online model support. RecStore maintains internal statistics and data structures aimed at providing efficient incremental updates to the recommender model, while employing an adaptive strategy for internal maintenance and load shedding to realize a balance between efficiency in updates or query processing based on system workloads. RecStore is also extensible, supporting a declarative syntax for defining recommender models. The efficacy of RecStore is demonstrated by providing the implementation details of three state-of-the-art collaborative filtering models. We provide an extensive experimental evaluation of a prototype of RecStore, built inside the storage engine of PostgreSQL, using a real-life recommender system workload.

UR - http://www.scopus.com/inward/record.url?scp=84863528189&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84863528189&partnerID=8YFLogxK

U2 - 10.1145/2247596.2247608

DO - 10.1145/2247596.2247608

M3 - Conference contribution

AN - SCOPUS:84863528189

SN - 9781450307901

SP - 86

EP - 96

BT - Advances in Database Technology - EDBT 2012

ER -