Subjective a priori grouping of tropical rain forest species for growth prediction may be unreliable because 1) there may be hundreds of species, many comparatively uncommon, the ecology of which may not be well known, 2) species within the same genus, may have significantly different growth patterns, and 3) growth rate may not provide a reliable indication of mortality. Growth models can retain the species identity of each simulated tree, but some aggregation is necessary to enable estimation of increment and mortality functions. An objective approach aggregated 100 rain forest tree species into ten groups to enable efficient estimation of mortality functions. This strategy provided better predictions than a previous subjective grouping. Annual survival probabilities were predicted from tree size, stand density and site quality using a logistic equation fitted by maximum likelihood estimation. Additional species with insufficient data for analysis were subjectively assigned to these ten equations. Several strategies were investigated; the best approach for these species seemed to be to employ the equation which served the greatest number of species. The increment pattern did not provide a good basis for assigning such species to equations, and this suggests that different groupings may be necessary to model the various components of tree growth.

Completely parallel object recognition is NP-complete. Achieving a recognizer with feasible complexity requires a compromise between parallel and sequential processing where a system selectively focuses on parts of a given image, one after another. Successive fixations are generated to sample the image and these samples are processed and abstracted to generate a temporal context in which results are integrated over time. A computational model based on a partially recurrent feedforward network is proposed and made credible by testing on the real-world problem of recognition of handwritten digits with encouraging results. 1 INTRODUCTION For all-parallel bottom-up recognition, allocating one separate unit for each possible feature combination, i.e., conjunctive encoding, implies combinatorial explosion. It has been shown that completely parallel, bottom-up visual object recognition is NP-complete (Tsotsos, 1990). By exchanging space with time, systems with much less complexity may be des...

. An analysis is given of the computational properties of Fisher's method of scoring for maximizing likelihoods and solving estimating equations based on quasi-likelihoods. Consistent estimation of the true parameter vector is shown to be important if a fast rate of convergence is to be achieved, but if this condition is met then the algorithm is very attractive. This link between the performance of the scoring algorithm and the adequacey of the underlying problem modelling is stressed. The effect of linear constraints on performance is discussed, and examples of likelihood and quasi-likelihood calculations are presented. 1. Introduction Two basic paradigms play important roles in the material developed in this paper. These are: (1) Newton's method for function minimization, and (2) the method of maximum likelihood for parameter estimation in data analysis problems. The main aim is to examine aspects of the structure and performance of Fisher's method of scoring, a minimization techniq...

This work considers probability models for partitions of a set of n elements using a predictive approach, i.e., models that are specified in terms of the conditional probability of either joining an already existing cluster or forming a new one. The inherent structure can be motivated by resorting to hierarchical models of either parametric or nonparametric nature. Parametric examples include the product partition models (PPMs) and the model-based approach of Dasgupta and Raftery (1998), while nonparametric alternatives include the Dirichlet Process, and more generally, the Species Sampling Models (SSMs). Under exchangeability, PPMs and SSMs induce the same type of partition structure. The methods are discussed in the context of outlier detection in normal linear regression models and of (univariate) density estimation.

INTRODUCTION Two-variable and multivariate analysis are important steps in fitting relationships for use in systems models. Many statistical packages for use on personal computers are available with regression capabilities, and there is great variation in range of capabilities, ease of use and cost. This is a personal overview of statistical packages used by the author, emphasizing the utility of the package for fitting curves to data using linear regression. It is not a comprehensive review, and does not consider expensive packages such as SPSS, SAS, and S-plus. Instead, it looks mainly at the free or cheap packages that do not require an annual license fee. Some basic concepts in regression analysis are first introduced, and then a number of packages with regression capabilities are reviewed -- specifically Excel, CurveExpert, GLIM, ARC and ViSta. 2. SOME BASIC CONCEPTS IN STATISTICS Let's begin by re-examining the principles underlying curve fitting with regression analysis. Fig

An analysis was conducted of 27,982 deaths among 106,020 persons employed at four Federal nuclear plants in Oak Ridge, Tennessee between 1943 and 1984. The main objectives were to extend the evaluation of the health effects of employment in the nuclear industry in Oak Ridge to include most workers who were omitted from earlier studies; to compare the mortality experience among the facilities; to address methodological problems that occur when individuals employed at more than one facility are included in the analysis; and to conduct dose-response analyses for those individuals with potential exposure to external radiation. All cause mortality and all cancer mortality were in close agreement with national rates. The only notable excesses occurred for white males for lung cancer---standardized mortality ratio (SMR) = 1.18, 1849 deaths--- and nonmalignant respiratory disease (SMR = 1.12, 1568 deaths). A more detailed analysis revealed substantial differences in death rates among workers a...

A two-stage model predicts recruitment of the 100 species which account for 97% of all recruitment observed on 217 permanent sample plots in the tropical rainforest of north Queensland. The first stage predicts the probability of the occurrence of any recruitment from stand basal area and the presence of that species in the existing stand. These probabilities can be implemented stochastically, or deterministically by summing for each species until unity is reached, recruitment initiated and the accumulated probability reset. The second stage indicates the expected amount of recruitment, given that it is known to occur, and employs stand basal area, the relative number of trees of that species in the stand, and site quality. This approach is easily implemented and provides good results.

Abstract not found

One of the objectives of the HUman MicroNucleus (HUMN) project is to identify the methodological variables that have an important impact on micronucleus (MN) or micronucleated (MNed) cell frequencies measured in human lymphocytes using the cytokinesis-block micronucleus assay. In a previous study we had shown that the scoring criteria used were likely to be an important variable. To determine the extent of residual variation when laboratories scored cells from the same cultures using the same set of standard scoring criteria, an inter-laboratory slide-scoring exercise was performed among 34 laboratories from 21 countries with a total of 51 slide scorers involved. The results of this study show that even under these optimized conditions there is a great variation in the MN frequency or MNed cell frequency obtained by individual laboratories and scorers. All laboratories ranked correctly the MNed cell frequency in cells from cultures that were unirradiated, or exposed to 1 or 2 Gy of gamma rays. The study also estimated that the intra-scorer median coefficient of variation for duplicate MNed cell frequency scores is 29% for unexposed cultures and 14 and 11% for cells exposed to 1 and 2 Gy, respectively. These values can be used as a standard for quality or acceptability of data in future studies. Using a Poisson regression model it was estimated that radiation dose explained 67% of the variance, while staining method, cell sample, laboratory, and covariance explained 0.6, 0.3, 6.5, and 25.6% of the variance, respectively, leaving only 3.1% of the variance unexplained. As part of this exercise, nucleoplasmic bridges were also estimated by the laboratories; however, inexperience in the use of this biomarker of chromosome rearrangement was reflected in the much greater heter...

Description Nonparametric maximum likelihood estimation or Gaussian quadrature for overdispersed