Describes a test collection that was developed as a multi-purpose testbed for experiments on the Web in distributed information retrieval, hyperlink algorithms, and conventional ad hoc retrieval. Discusses inter-server connectivity, integrity of server holdings, inclusion of documents related to a wide spread of likely queries, and distribution of…

Discusses the use of text retrieval methods based on probabilistic models with Chinese language material, which are modeled on the Okapi information retrieval system. Topics include system architecture, test collections, weighting functions, and algorithms. (Author/LRW)

Presents a genetic relevance optimization process performed in an information retrieval system that uses genetic techniques for solving multimodal problems (niching) and query reformulation techniques. Explains that the niching technique allows the process to reach different relevance regions of the document space, and that query reformulations…

General algorithms are presented that can be used for optimizing matrix trace functions subject to certain constraints on the parameters. The parameter set that minimizes the majorizing function also decreases the matrix trace function, providing a monotonically convergent algorithm for minimizing the matrix trace function iteratively. (SLD)

Analyzes the computational complexity of optimal binary tree pruning for tree-structured vector quantization. Topics discussed include the combinatorial nature of the optimization problem; the complexity of optimal tree pruning; and finding a minimal size pruned tree. (11 references) (LRW)

It is shown that it is not easy to apply the EM algorithm to latent class models in the general case with equality constraints because a nonlinear equation has to be solved. A simpler condition is given in which the EM algorithm can be easily applied. (SLD)

Describes a string-searching algorithm for multiple patterns in a text string; explains the construction of a pattern-matching machine; presents a theoretical analysis and empirical evidence that supports the sublinearity of the algorithm; and compares this algorithm with the Boyer-Moore algorithm for a single pattern. (Contains 10 references.)…

Discusses the minimum average codeword length coding under the constraint that the codewords are monotonically nondecreasing in length. Bounds on the average length of an optimal monotonic code are derived, and sufficient conditions are given such that algorithms for optimal alphabetic codes can be used to find the optimal monotonic code. (six…

Describes the maximum likelihood (ML) estimation of mean and covariance structure models when data are missing. Describes expectation maximization (EM), generalized expectation maximization, Fletcher-Powell, and Fisher-scoring algorithms for parameter estimation and shows how software can be used to implement each algorithm. (Author/SLD)

Presents a three-way, three-mode extension of the two-way, two-mode hierarchical classes model of P. De Boeck and S. Rosenberg (1998) for the analysis of individual differences in binary object x attribute arrays. Illustrates the model with data on psychiatric diagnosis and discusses the relation between the model and other extant models. (SLD)

Introduces a new algorithm for obtaining exact person fit indexes for the Rasch model. The algorithm realizes most tests for a general family of alternative hypotheses, including tests concerning differential item functioning. The method is also used as a goodness-of-fit test in some circumstances. Simulated examples and an empirical investigation…

The classification capabilities of the one-dimensional Kohonen neural network (T. Kohonen, 1995) were compared with those of two partitioning and three hierarchical cluster methods in 2,580 data sets with known cluster structure. Overall, the performance of the Kohonen networks was similar to, or better than, that of the others. Implications for…

Discussion of semantic networks and ranked retrieval focuses on two models, the semantic network model with spreading activation and the vector space model with dot product. Suggests a formal method to analyze the two models in terms of their relative performance in the same universe of objects. (Author/LRW)

Studied the calculating ability used by a low IQ savant to identify prime numbers in two experiments comparing him to control subjects, one involving reaction time and the other involving inspection time. Concludes that this individual uses a complex computational algorithm to identify primes and discusses the apparent contradiction of his low IQ.…

Discussion of the classification of documents from bibliographic databases focuses on a method of vectorizing reference documents from LISA (Library and Information Science Abstracts) which permits their topological organization using Kohonen's algorithm. Analyzes possibilities of this type of neural network with respect to the development of…