Multiversion information retrieval systems and feedback with mechanism of selection

Frants, Valery I.; Shapiro, Jacob; Voiskunskii, Vladimir G.

doi:10.1002/(sici)1097-4571(199301)44:1<19::aid-asi3>3.0.co;2-1

Cited by 11 publications

(8 citation statements)

References 7 publications

(8 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…We note, incidentally, that comparison of search results is also based on analysis of the sign of the difference of obtained values of the corresponding criterion. In fact this, as well as the existence of the order preservation property for the measure 12, allowed criterion r2 / n to be used for comparison of search results in Frants et al (1993). Now we will show that the series of constructed singlevalued measures, namely: 12, 13, 14, and 15, have the order preservation property.…”

Section: Order Preservation Propertymentioning

confidence: 88%

“…Nevertheless, such criteria do exist. One of them, r2/ n (where r is the numker of relevant documents in the output, and n is the number of documents in the output) was used in Frants, Shapiro, and Voiskunskii ( 1993) for construction of a selection mechanism in a feedback process. The analytic notation of this criterion is only a more convenient representation of single-valued measure ]2 =~P for a case of comparison of search results.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Evaluation of search results: A new approach

Voiskunskii

1997

J. Am. Soc. Inf. Sci.

Self Cite

View full text Add to dashboard Cite

In the present work, a new approach to construction of single‐valued measures for evaluation of search results is considered in detail. This approach is based first on the formation of certain vectors, and second on the use of functions allowing the evaluation of their closeness. Realization of this approach leads to construction of new, as well as widely‐known, single‐valued measures. A series of constructed measures has a property important in practice, leading to the creation of a mechanism which would allow the selection of the best search method (from a set of available methods) for every query. In this article, a general statement of this property, called the order preservation property, is given, and all considered single‐valued measures are analyzed for the existence of this property. Finally, problems of the legitimacy of the use of single‐valued measures in various specific situations are discussed in the article, and an approach to the solution of these problems is suggested. © 1997 John Wiley & Sons, Inc.

show abstract

Section: Order Preservation Propertymentioning

confidence: 88%

Section: Introductionmentioning

confidence: 99%

Evaluation of search results: A new approach

Voiskunskii

1997

J. Am. Soc. Inf. Sci.

Self Cite

View full text Add to dashboard Cite

show abstract

“…In the second step, each token is stemmed into its root form; for example, a verb in its past is transformed into its root form, and a noun in its plural form is transformed into its singular form. Words which function only grammatically with regardless of a content are called stop words [5], and they correspond to articles, conjunctions, or pronouns. In the third step, stop words are removed for processing documents more efficiently and reliably for text categorization.…”

Section: Strategies Of Encoding Documentsmentioning

confidence: 99%

Neural Text Categorizer for Exclusive Text Categorization

Jo¹

2008

Journal of Information Processing Systems

View full text Add to dashboard Cite

This research proposes a new neural network for text categorization which uses alternative representations of documents to numerical vectors. Since the proposed neural network is intended originally only for text categorization, it is called NTC (Neural Text Categorizer) in this research. Numerical vectors representing documents for tasks of text mining have inherently two main problems: huge dimensionality and sparse distribution. Although many various feature selection methods are developed to address the first problem, the reduced dimension remains still large. If the dimension is reduced excessively by a feature selection method, robustness of text categorization is degraded. Even if SVM (Support Vector Machine) is tolerable to huge dimensionality, it is not so to the second problem. The goal of this research is to address the two problems at same time by proposing a new representation of documents and a new neural network using the representation for its input vector.

show abstract

“…The results of the experiment were described in [1]. For the benefit of the reader, we describe it in the Appendix.…”

Section: Information Need (In) and Information Retrieval (Ir)mentioning

confidence: 99%

“…In 1993, we introduced the notion of multiversion IR systems capable of performing optimal searches for each individual user [1]. The main emphasis in the paper was to demonstrate a simple and effective mechanism for choosing the system's best state for an individual user; in particular, in the case of an online system.…”

Section: Introductionmentioning

confidence: 99%

Optimal search available to an individual user

Frants

Shapiro

Voiskunskii

1996

Journal of Information Science

Self Cite

View full text Add to dashboard Cite

The paper describes an approach to the design and implementation of an information retrieval (IR) system capable of providing an optimal search for each individual user. On the basis of the properties of information need, the paper analyzes the peculiarities of optimal search in documentary IR systems. The problems arising in implementing optimal search in such a system are considered and concrete solutions of some of these problems are proposed. In particular, the paper describes a simple and convenient method of selecting the 'best result' and proposes an algorithm for organizing optimal search in the systems realizing SDI (selective dissemination of information). This algorithm can be utilized, without substantial modifications, in many systems capable of using different search methods.

show abstract

Multiversion information retrieval systems and feedback with mechanism of selection

Cited by 11 publications

References 7 publications

Evaluation of search results: A new approach

Evaluation of search results: A new approach

Neural Text Categorizer for Exclusive Text Categorization

Optimal search available to an individual user

Contact Info

Product

Resources

About