Text retrieval and the relational model

Macleod, Ian A.

doi:10.1002/(sici)1097-4571(199104)42:3<155::aid-asi1>3.0.co;2-h

Cited by 11 publications

(7 citation statements)

References 6 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…Examples of early work include [13,10,6]; key references on runtime efficiency are discussed in a 2005 survey [3]. Perhaps the first 'real' SQL results for IR queries were presented in [7], but at excessively high costs in terms of the hardware required.…”

Section: Background and Related Workmentioning

confidence: 99%

Old dogs are great at new tricks

Mühleisen

Samar

Lin

et al. 2014

Proceedings of the 37th International ACM SIGIR Conference on Research &Amp; Development in Information Retrieval

View full text Add to dashboard Cite

We make the suggestion that instead of implementing custom index structures and query evaluation algorithms, IR researchers should simply store document representations in a column-oriented relational database and implement ranking models using SQL. For rapid prototyping, this is particularly advantageous since researchers can explore new scoring functions and features by simply issuing SQL queries, without needing to write imperative code. We demonstrate the feasibility of this approach by an implementation of conjunctive BM25 using two modern column stores. Experiments on a web collection show that a retrieval engine built in this manner achieves effectiveness and efficiency on par with custom-built retrieval engines, but provides many additional advantages, including cleaner query semantics, a simpler architecture, built-in support for error analysis, and the ability to exploit advances in database technology "for free".

show abstract

Section: Background and Related Workmentioning

confidence: 99%

Old dogs are great at new tricks

Mühleisen

Samar

Lin

et al. 2014

Proceedings of the 37th International ACM SIGIR Conference on Research &Amp; Development in Information Retrieval

View full text Add to dashboard Cite

show abstract

“…Information content requires constant monitoring & handling, especially in some cases where inappropriate access of the information may result in penalties, commercial losses, identity theft, privacy invasion, or unauthorized access by an identity or many identities [10]. Figure 7 Categorizes sensitivity on the basis of three factors -low, medium and high.…”

Section: Sensitivitymentioning

confidence: 99%

Security Issues and Accuracy Concerns in Information Retrieval Process

Gautam¹,

Jain²,

Pushkar³

et al. 2013

IJCA

View full text Add to dashboard Cite

An approach towards Information retrieval process security issues & accuracy of the retrieved data from the user perspective, are the major concerns of this paper. Implanted work includes demonstration of the retrieval modes, process flow of information retrieval system and risk analysis of the retrieved data. Much work has been done to highlight the accuracy problem in the retrieved content where an accuracy formula focuses on the calculation of the accuracy percentage in order to generate an transparent approach towards the information retrieval process.

show abstract

“…Diese Ausführungen wurden von Macleod [13] und Blair [3] fortgeführt. Aus diesen Artikeln wird aber auch ersichtlich, wie komplex die Anfragen werden können.…”

Section: Realisierungsalternativenunclassified

“…Die Verwendung des relationalen Modells für Textretrievalanwendungen beschreibt bereits die frühe Arbeit von Macleod [12], ohne jedoch auf Performanzaspekte einzugehen [3,13]. Frühe Studien von Schek und Stonebraker [18,20] machen auf die Unzulänglichkeiten des relationalen Modells bei der Verwaltung von Textdaten aufmerksam [13].…”

Section: Verwandte Arbeitenunclassified

See 1 more Smart Citation

Text-Retrieval mit einem relationalen Datenbank-Management-System

Kalinski

1999

Informatik - Forschung und Entwicklung

View full text Add to dashboard Cite

Zusammenfassung.Wir untersuchen Techniken des TextRetrievals mit relationalen Datenbanken. Den Tests zum Booleschen Retrieval liegen reale Bibliotheksdaten und Bibliothekarsanfragen zugrunde. Es werden drei Realisierungen mittels SQL miteinander verglichen, von denen der Nested Loops Join mit vorsortierter Wortfolge am besten abschneidet. Dessen Effizienz beruht auf den signifikant unterschiedlichen Vorkommenshäufigkeiten der Anfragewörter. Um diese Beobachtung auch bei einer Implementierung mit komprimierten Bitlisten effizienzsteigernd zu nutzen, wird eine zusätzliche Zugriffsstruktur für Lookup-Operationen in den Bitlisten aufgebaut. Im Gegensatz zum Booleschen Retrieval zeigt sich beim Retrieval mit Ranking, daß eine Ergebnisberechnung mittels eines geschlossenen SQL-Ausdrucks einer Lösung, die im Anwendungsprogramm die invertierten Listen durchläuft und zu einem Ranking kombiniert, hinsichtlich der Effizienz unterlegen ist. Als Grundlage für Abbruchheuristiken wird hier außerdem eine Indexstruktur aufgebaut, die den Zugriff auf invertierte Listen sortiert nach Within-DocumentFrequencies erlaubt.Abstract. Our paper discusses techniques for text retrieval with relational database systems. The experiments for Boolean retrieval, which are based on real-life library data and librarians' queries, compare three alternative SQL realizations. It turns out that a nested loops join with pre-sorted word order is most efficient because keywords have significantly different frequency values. This observation can also be used to enhance a realization based on compressed bit lists. An additional access structure is needed that supports lookup operations within bit sequences. In the context of retrieval with ranking, however, results are best computed under the regime of the application program. A cursor-based traversal and combination of inverted lists is more efficient than a closed SQL query. Further performance improvements benefit from an additional index which supports access to inverted lists in order of decreasing within-document-frequencies.

show abstract

Text retrieval and the relational model

Cited by 11 publications

References 6 publications

Old dogs are great at new tricks

Old dogs are great at new tricks

Security Issues and Accuracy Concerns in Information Retrieval Process

Text-Retrieval mit einem relationalen Datenbank-Management-System

Contact Info

Product

Resources

About