Extreme-value statistics is often used to estimate so-called return values (actually related to quantiles) for environmental quantities like wind speed or wave height. A basic method for estimation is the method of block maxima which consists in partitioning observations in blocks, where maxima from each block could be considered independent.…

Coefficients of correlation and concordance are defined for profiles consisting of ensembles of partially ordered sets, by extending Kendall's rank order correlation (tau) and Kendall's coefficient of concordance (W) to cases where subjects are asked to rank order, or to choose among, the given alternatives in multiple sets of items. (Author/PN)

Under certain conditions, a validity Coefficient of Effective Length (CEL) can produce highly misleading results. A modified coefficent is suggested for use when empirical studies indicate that underlying assumptions have been violated. (Author/BW)

A more general method for calculating the Kappa measure of nominal rating agreement among multiple raters is presented. It can be used across a broad range of rating designs, including those in which raters vary with respect to their base rates and how many subjects they rate in common. (Author/BW)

Four separate approaches to the problem of assessing variation in categorical data are developed, each of which results in an identical index of dispersion. An additional index, related to a measure of entropy for categorical data, is mentioned and some applications are discussed. (Author/RL)

Power and stability of Type I error rates are investigated for the Box-Scheffe test of homogeneity of variance with varying subsample sizes under conditions of normality and nonnormality. The test is shown to be robust to violation of the normality assumption when sampling is from a leptokurtic population. (Author/GK)

Guilford's average interrater correlation coefficient is shown to be related to the Friedman Rank Sum statistic. Under the null hypothesis of zero correlation, the resultant distribution is known and the hypothesis can be tested. Large sample and tied score cases are also considered. An example from Guilford (1954) is presented. (Author)

The F-distribution approximation suggested by Dixon was investigated at various combinations of alpha and degrees of freedom. Tabled values were compared with values computed utilizing the suggested formula. (Author/GSK)

The value of the phi coefficient cannot, in general, be uniquely determined from the G index. G is more suitable than the phi coefficient as a measure of similarity between individuals, because phi is too unstable over item reflection. (Author/RL)

Discusses sequential patterns, data mining, knowledge acquisition, and fuzzy sequential patterns described by natural language. Proposes a fuzzy data mining technique to discover fuzzy sequential patterns by using the simple partition method which allows the linguistic interpretation of each fuzzy set to be easily obtained. (Author/LRW)

Presents a general method to construct ordered similarity measures in information retrieval based on classical similarity measures for ordinary sets. Describes a test of some of these measures in an information retrieval system that extracted ranked document sets and discuses the practical usability of the ordered similarity measures. (Author/LRW)

Reports data on the application of Bradford's law to several classical bibliographies, for which the underlying law of Leimkuhler and various Bradford groupings are determined. The Leimkuhler law is extended to be applicable to incomplete bibliographies and to the estimation of complete bibliographies from incomplete bibliographies. Formulas and…

Presents a class of distortion controlled vector quantizers that are capable of compressing images so they comply with certain distortion requirements. Highlights include tree-structured vector quantizers; multiresolution vector quantization; error coding vector quantizer; error coding multiresolution algorithm; and Huffman coding of the quad-tree…

Describes the development of new statistical formulas for identifying two- and three-character words in Chinese text by performing stepwise logistic regression using a sample of sentences that had been manually segmented. Concludes that the new contextual information formulas are substantially better than the mutual information formula.…

Argues that informetrics needs a rigorous mathematical basis and that informetric processes can be adequately describes within a standard probabilistic framework with stochastic modeling offering a more productive approach than other methods. Highlights include Lotka's Law; robustness and multiauthorship; and resilience and the random component.…