Myaeng, S., & Lopez-Lopez, A. (1992). Conceptual graph matching: A flexible algorithm and experiments. Experimental and Theoretical Artificial Intelligence, 4, 107--126.  Home/Search   Document Not in Database   Summary   Related Articles   Check

....graph [2,4] in this case neither description nor measure of their similarity is obtained. Some other, more general methods, do measure the similarity between two conceptual graphs, but they typically describe this similarity as the set of all their common elements allowing duplicated information [3,6,7]. Yet other methods are focused on question answering [12] these methods allow a flexible matching of the graphs, but they do not compute any similarity measure. Work done under partial support of CONACyT, CGEPI IPN, and SNI, Mexico. The method we propose is general but flexible. First, it ....

....the algorithm to find the overlaps. Given two conceptual graphs G 1 and G 2 , the goal is to find all their overlaps. Our algorithm works in two stages. At the first stage, all similarities (correspondences) between the conceptual graphs are found, i.e. a kind of the product graph is constructed [6]. The product graph P Recall that projection map and thus the projection for a given pair v , u is not unique. Matching Similarity measure overlaps Domain Knowledge User conceptual graphs Similarity description Similarity measure Figure 3. Comparison of conceptual graphs expresses the ....

[Article contains additional citation context not shown here]

Myaeng, Sung H., and Aurelio Lpez-Lpez (1992), "Conceptual Graph Matching: a Flexible Algorithm and Experiments", Journal of Experimental and Theoretical Artificial Intell i - gence, Vol. 4, 1992.

....to noun phrases, can be used to compute the probability that a document is relevant given a certain profile or request. ffl Conceptual graphs (see e.g. Sow84] can be used for the modeling of noun phrases. After this has been done, known similarity measures for conceptual graphs (see e.g. MLL92] and [Mah93] can be used to define a similarity measure between noun phrases and thus for similarity measures between documents and a query or profile. ffl A Phrase Space, an adaptation of the Vector Space Model to noun phrases, can be developed to obtain a measure of similarity between vectors ....

S.H. Myaeng and A. Lopez-Lopez. Conceptual graph matching: a flexible algorithm and experiments. Journal of Experimental Theoretical Artificial Intelligence, 4:107--126, 1992.

....measure for flowsheets should be based not only on the local structure of the objects but on the relational structure existing between them as well. Several approaches were made in this field including: analogical reasoning (Holyoak Thagard 1989, Falkenhainer et al. 1989) conceptual graphs (Myaeng Lopez Lopez 1992, Maher 1993) and object oriented representation (Bisson 1995) In CBR the mentioned FABEL project investigates several approaches to structural similarity based on: graph theory, term based representation, gestalt indices and psychological theories of perception (Voss et al. 1994a) The ....

Myaeng S., Lopez-Lopez A. (1992). Conceptual graph matching: a flexible algorithm and experiments. J.Expt, Theor. Artif. Intell., vol.4, pp.107-126.

.... improvement of the backtracking method was then presented by Ullman, who introduced a refinement method which reduces the search space of the backtracking procedure remarkably [Ull76] For an overview of the publications on graph isomorphism see [Gat79, RC77] More recent work is described in [MLL92, FFG90] where the graph isomorphism problem was reduced to the problem of clique detection by constructing an association graph for all the possible vertex mappings. And most recently, a network based approach to graph matching has been proposed by the authors [MB95b] So far, we have only considered the ....

S.H. Myaeng and A. Lopez-Lopez. Conceptual graph matching: a flexible algorithm and experiments. Journal of Experimental and Theoretical Artificial Intelligence, 4:107--126, April 1992.

....essentially look for maximal common subgraphs between the two input graphs. The similarity of the graphs increases as the size of this common subgraph increases. Examples of such an approach include Leishman s work on analogy [10] and the algorithms of McGregor [12] and Myaeng and Lopez Lopez [15]. Leishman s work is unusual in that she considers evaluation functions that chose between competing minimal common generalizations. The problem with most algorithms reported is that they tend not to consider exactly what form of similarity they are trying to measure, but instead implicitly ....

....makes use of the observation that the interest of P itself forms an upper bound on the interest of any common generalization. 3.2 Forming a Product Graph The first step is to create a product graph. It is equivalent to the association graph formed as the first step in the algorithm presented in [15]. Essentially the cartesian product of the nodes and arcs is formed, where the pairs are combined using the join function. Matched node pairs or arc pairs with empty common generalizations (i.e. are removed. We assume that the labels on nodes (concepts) and arcs (relations) can be treated as if ....

Sung H. Myaeng and Aurelio Lopez-Lopez. Conceptual graph matching: a flexible algorithm and experiments. Journal of Experimental and Theoretical Artificial Intelligence, 4:107--126, April--June 1992.

....Index Lithoid1 Lithoid2 Common terms Worst Case Association Index Figure 10: Growth of Association Index for Cranfield Collection. 3.7.1 Association Graphs The association language expresses the nodes that can appear in an association index. The nodes are more structured than those defined in [MLL92] for association graphs, which makes our association language a special case of the general node language for association graphs. A more enriched association language could be defined, for instance including the complete Cartesian product of the user and source lithoids. However, for our purpose, ....

....the complete Cartesian product of the user and source lithoids. However, for our purpose, as sketched by the criteria in the introduction, the defined association language is a fair approach. The association relation is a special case of the set of edges defined for association graphs in [MLL92] This is a natural consequence of the fact that the association language is a special case of the association nodes for association graphs. 3.7.2 Galois Lattices Johannesson describes the use of Galois lattices for supporting schema integration in [Joh95] The Galois lattices are used to ....

[Article contains additional citation context not shown here]

S.H. Myaeng and A. Lopez-Lopez. Conceptual graph matching: a flexible algorithm and experiments. Journal of Experimental Theoretical Artificial Intelligence, 4:107--126, 1992.

....on concept nodes C 2 and relation nodes R 2 from G 2 . type matching: in this phase the result of node matching is further refined based on node types. n m i 1 k t t t 1 2 3 2 i .name = AA .name = BB .name = 1 i 2 i 2 t 3 . name t Figure 13: Restricting attribute values This is in line with [MLL92] The result of the first phase is called an association graph. The association graph is the basis for similarity computation. This graph GA = hCA ; RA ; EA i is constructed as follows. The nodes in GA are given by CA = C 1 Theta C 2 and RA = R 1 Theta R 2 . The edges in GA are constructed ....

S.H. Myaeng and A. Lopez-Lopez. Conceptual graph matching: a flexible algorithm and experiments. Journal of Experimental Theoretical Artificial Intelligence, 4:107--126, 1992.

.... In the simplest case, where the Conceptual Graphs of the two cases to be compared are isomorphic, the similarity of the two cases can perhaps be defined as a function of the similarity of the aligned 2 concepts and relations in the two graphs (this is suggested as a similarity metric for CBR in [8]) Even within this very simplistic scheme for comparing structured representations, however, there are complications particularly with regard to defining a static similarity metric for individual concepts (see below) More importantly, however, it is hard to guarantee that two Conceptual Graphs ....

Sung H. Myaeng and Aurelio Lopez-Lopez. Conceptual graph matching: a flexible algorithm and experiments. Journal of Experimental and Theoretical Artificial Intelligence, 4:107--126, April--June 1992.

....can be used in data mining and or case based retrieval in object oriented databases. At present there is considerable growing interest in the use of structural representations, including research in: analogical reasoning (Holyoak Thagard 1989, Falkenhainer et al. 1989) conceptual graphs (Myaeng Lopez Lopez 1992, Maher 1993) and objectoriented representation (Bisson 1995) In case based reasoning the FABEL project investigates several approaches for defining the similarity of structures (Voss et al. 1994) In contrast to the most of these approaches, the proposed similarity measure is computationally ....

Myaeng S., Lopez-Lopez A. (1992). Conceptual graph matching: a flexible algorithm and experiments. J.Expt, Theor. Artif. Intell., vol.4, pp.107-126.

....list of documents as the third and final output of the system. Using the techniques necessary to model plausible inferences with CGs (Myaeng and Khoo, 1992) this module computes the degree to which the topic statement CG is covered by the CGs in the document (see Myaeng and Liddy (1993) and Myaeng Lopez lopez (1992) for details) While the most obvious strength of the CG approach is its ability to enhance precision by exploiting the structure of the CGs and the semantics of relations in document and topic statement CGs, and by attempting to meet the specific semantic constraints of topic statements, we also ....

....the original motivation was to represent documents and topic statements at more conceptual level using RIT codes, we are also testing the effectiveness of RIT based term expansion in IR environments. Using the scheme we have developed for term clustering using contextual information in the corpus (Myaeng Li, 1992), we have three methods to evaluate: RIT based expansion, term cluster based expansion, and a combination of the two so that we can eliminate the problem of using a general purpose thesaurus and the errors made by the term clustering method. For TIPSTER evaluation, we have submitted two sets of ....

Myaeng, S. H. & Lopez-Lopez, A. (1992). Conceptual graph matching: A flexible algorithm and experiments.

No context found.

Myaeng, S., & Lopez-Lopez, A. (1992). Conceptual graph matching: A flexible algorithm and experiments. Experimental and Theoretical Artificial Intelligence, 4, 107--126.

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