The purpose of this methodological investigation was to study the relative magnitude of the various errors in vertical ridge measurements from orthopantomograms of edentulous patients.The study was divided into two parts. Part I and Part II. Part I included eleven patients. Two standard orthopantomograms were taken of each patient. Part II included five ofthe eleven patients in Part I. Here, a third X-ray taken with the mouth open was added to the X-rays from Part I. The X-rays were traced by three dentists.Twelve vertical measurements in the maxilla and in the mandible were made. The variance produced by the patients (morphological variance) and those produced by the dentist, the X-ray, the interaction patient-dentist, the interaction dentist-X-ray and the random variance (methodological variances) were separately estimated on the basis of a three-way mixed analysis of variance model.In Part I, on the average, the sum ofthe methodological variances constituted 1 -7 % and 2-5 % of the total variance in the maxilla and in the mandible respectively. The X-ray produced the greatest single variance component, averaging M % in both jaws. Although generally statistically significant, the remaining methodological variances were of low magnitude. The interaction patient-dentist, however, was not significant in most cases.In Part II, on the average, the sum of the methodological variances increased to 38-8% in the maxilla and 4-6% in the mandible, mainly caused by an increase in the variance produced by the X-ray.
Various methods for contextual classification of multi-are determined from some simulated SPOT data. Three spectral scanner data have been developed during the last 15 years, replications of the experiment are carried out for each aiming at increased accuracy in classified images. The methods have for a large part been of four main types: 1) neighborhood-based clas-.cti of scene a utoration a a trames. The sification based on stochastic models for the classes over the scene and effects of parameter estimation from a training set are for the vectors given the classes; 2) simultaneous classification of all studied as well. pixels, using, e.g., Markov random-field models; 3) relaxation methods Further technical descriptions of the classification prothat iteratively modify posterior probabilities using information from cedures are available in SaebO et al. [5], where applicaan increasing neighborhood; and 4) methods using ordinary noncon-tions to real world data also are included. Further infortextual rules based on transformed data. In the present paper a selection of these methods is presented and compared using computer-gen-mation about the Monte Carlo study iS available in a erated data on different scenes. Spatial autocorrelation is present in technical report and can be obtained from the authors upon the data. Error rates are compared, and an attempt is made to char-request.acterize what kind of errors each particular method makes.Monte Carlo studies of the present type quickly become Keywords-Contextual classification, spatial autocorrelation, Monte both cumbersome and expensive; the methods typically Carlo study, remote sensing.demand much CPU time, and a lot of information needs to be condensed. These factors naturally limited the ex-
FlekkØy, K., Holme, I. & Mohn, E. Relation of number of different responses to group size and stimulus words in a discrete free‐association situation. Scand. J. Psychol., 1973, 14, 4–8.‐An exploration of the total number (D) of different responses given to a stimulus word in a single‐word, free‐association situation showed that (1) the first 50 Kent‐Rosanoff words (in Norwegian) were highly heterogeneous with regard to D; (2) was an asymptotic function of the number of subjects; and (3) the larger D for schizophrenics may be attributed to their higher frequency of individual responses. It was also indicated that D may not be a valid measure of the magnitude of the associative potential of the individual.
JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship. For more information about JSTOR, please contact support@jstor.org.. Biometrika Trust is collaborating with JSTOR to digitize, preserve and extend access to Biometrika. SUMMARYBased on a random sample from the N(6, a2) distribution, we find confidence bands with level 1a for exp (6 + ba2) for all b. The method is based on a technique discussed by Kanofsky (1968a, b). Using the likelihood ratio confidence region for (6, a2) (Spj0tvoll, 1972) as starting point we find that the probability that the confidence interval for a particular value of b covers the true value is fairly constant under variation of b, a and the sample size. This makes it possible to construct confidence intervals with levels approximately equal to a preassigned value. Choosing b = 2 we obtain an interval for the expectation of the log normal distribution. It has practically the same expected width as an approximate interval suggested by Cox (Land, 1972). The confidence band obtained by using the confidence region proposed by Mood (1950, p. 228) is compared with the likelihood ratio band and seems to have less favourable statistical properties.Some key words: Confidence band; Likelihood ratio; Log normal distribution; Measure of location.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.
hi@scite.ai
10624 S. Eastern Ave., Ste. A-614
Henderson, NV 89052, USA
Copyright © 2024 scite LLC. All rights reserved.
Made with 💙 for researchers
Part of the Research Solutions Family.