| VAN RIJSBERGEN, C. J. 1986. A non-classical logic for information retrieval. The Computer Journal 29, 6, 481--485. |
....[15, 10] Documents and queries are represented as Propositional Logic formulas. Given a document and a query represented by the propositional formulas d and q respectively, it is well known that the application of the notion of logical consequence to decide relevance, i.e. d j= q, is too strict [23]. The entailment d j= q simply tests whether or not each logical interpretation that makes d true makes also q true (i.e. each model of d is also a model of q) This is not in accordance with what we expect from an IR measure of relevance. Let us illustrate it through an example. Imagine two ....
C.J. van Rijsbergen. A non-classical logic for information retrieval. The Computer Journal, 29:481485, 1986.
....this model heavily depends on good approximations of the number of relevant documents in the result set. We will present and evaluate three different methods in the next sections. 3. RELEVANT DOCUMENTS IN THE RESULT SET As usual in modern IR, we view information retrieval as uncertain inference [21] and consider the probability Pr(q d) that document d logically implies the query q. The relationship between this probability of inference and the probability Pr(rel q,d) of d being relevant to q is modelled with a query specific function [0, 1] Pr(rel q,d) f (Pr(q d) A query q in this ....
....library, and the retrieved documents are collected. With this technique, a reasonable accurate resource description can be constructed. 3.1 Method 1: Recall precision function This method is already described in detail in [9] here we use it as a baseline for the other methods. According to [21], we can compute the probability of relevance as Pr(q d) Pr(rel (q d) Pr( q d) If we assume Pr(rel (q d) 0, we obtain Pr(q d) Thus, the relationship between the probabilities Pr(rel q,d) and Pr(q d) can be modelled with the linear function [0, 1] f (x) Pr(rel q d) ....
C. J. van Rijsbergen. A non-classical logic for information retrieval. The Computer Journal, 29(6):481--485, 1986.
....a vector (E i (0, q) E i (1, q) EC i(n, q) T. The dispatcher then uses these values to compute the optimum selection vector with one of the algorithms described in section 2.2. Computing relevance probabilities As usual in modern IR, we view information retrieval as uncertain inference [17]. Thus, for a query q, the probability Pr(rel[q, d) of a document being relevant can be formulated based on the probability of the uncertain implication Pr(q d) with the query specific function Pr(rellq , d) f(Pr(q d) f: 0,1] 0,1] For computing the implication probability Pr(q d) ....
....f: 0, 1] 0, 1] In this rest of this chapter, two different functions namely a linear function and a logistic function are described. Both functions are evaluated in section 5.4. 3. 1 Linear function The probability Pr(rellq , d) can be formulated based on the formula of total probabilities [17]: Pr(rellq ,d) Pr(rellq d) Pr(q d) q Pr(rell(q d) Pr( q d) Pr(rellq d) Pr(q d) Thus, the relationship between the probabilities Pr(rellq, d) and Pr(q d) the score) can be modelled as the linear function f: 0,1] 0,1] f(x) Pr(rellq d) x. If no relevance data is ....
C. J. van Rijsbergen. A non-classical logic for information retrieval. The Computer Journal, 29(6):481485, 1986.
No context found.
van Rijsbergen, C. J. A non-classical logic for information retrieval. The Computer Journal 29, 6 (1986), 481--485.
No context found.
C.J. van Rijsbergen. A non-classical logic for information retrieval. The Computer Journal, 29:481-485, 1986.
No context found.
VAN RIJSBERGEN, C. J. 1986. A non-classical logic for information retrieval. The Computer Journal 29, 6, 481--485.
No context found.
C. J. van Rijsbergen (1986). \A Non-classical Logic for Information Retrieval," The Computer Journal, 29(6).
No context found.
C. J. van Rijsbergen. A non-classical logic for information retrieval. The Computer Journal, 29(6), 1986.
No context found.
Keith van Rijsbergen. A non-classical logic for information retrieval. The Computer Journal, 29:481-- 485, 1986. 307
No context found.
C. J. van Rijsbergen. A non-classical logic for information retrieval. The Computer Journal, 29(6), 1986.
No context found.
van Rijsbergen, C. J. (1986). A non-classical logic for information retrieval. The Computer Journal, 29(6).
No context found.
VAN RIJSBERGEN, C. J. 1986. A non-classical logic for information retrieval. The Computer Journal 29, 6, 481--485.
No context found.
van Rijsbergen, C. J. (1986). A non-classical logic for information retrieval. The Computer Journal, 29(6).
No context found.
C. J. van Rijsbergen. A non-classical logic for information retrieval. The Computer Journal, 29(6):481--485, 1986.
No context found.
van Rijsbergen, C. J. (1986). A non-classical logic for information retrieval. The Computer Journal, 29(6):481--485.
No context found.
Keith van Rijsbergen. A non-classical logic for information retrieval. The Computer Journal, 29:481--485, 1986.
No context found.
C.J. van Rijsbergen. A non-classical logic for information retrieval. The Computer Journal, 29(6):481--485, 1986.
No context found.
C. J. van Rijsbergen. A non-classical logic for information retrieval. The Computer Journal, 29(6):481--485, 1986.
No context found.
C. J. Van Rijsbergen. A non-classical logic for information retrieval. The Computer Journal, 29:481--485, 1986.
No context found.
C.J. van Rijsbergen. A non-classical logic for information retrieval. The Computer Journal, 29:481485, 1986.
No context found.
van Rijsbergen, C. J. (1986) A non-classical logic for information retrieval. Comp. J., 29, 481--485.
No context found.
C. J. van Rijsbergen, A non-classical logic for information retrieval, Computer Journal, 29, pp. 481-485, 1986. 17
No context found.
K. van Rijsbergen. A non-classical logic for information retrieval. The Computer Journal, 29:481--485, 1986.
No context found.
C. J. van Rijsbergen. A non-classical logic for information retrieval. The Computer Journal, 29(6):481--485, 1986.
No context found.
C.J, van Rijsbergen, A non-classical logic for Information Retrieval, The Computer Journal, 29: 481-485, 1986.
First 50 documents
Online articles have much greater impact More about CiteSeer.IST Add search form to your site Submit documents Feedback
CiteSeer.IST - Copyright Penn State and NEC