A generalization of estimation method based on Bentler-Weeks model (1980) is discussed for some cases of arbitrary functional constraints on parameters, that is, linear constraints, arbitrary equality constraints and inequality constraints. The purpose of the present paper is to give concretely an extensive algorithm that, using both methods of penalty function and relaxation function, two cases with equality and inequality constraints can be solved by the similar way as Lee (1980). A numerical example is also presented to illustrate the proposed algorithm.
STD is transmitted by coitus. Hence, the spread of the disease will be influenced by sexual freedom in a society. The quantitative relationship between them is analyzed here. When αβμ' φ' < 1, the STD will subside eventually, while when αβμ' φ' ≥ 1, it is possible to spread explosively where μ'= ?? (σ2+μ2-μ), φ'= ?? (σ2+φ2-φ) μ and σ1 are the mean and standard deviation of p(k) φ and σ2 are the mean and standard deviation of q(k) α: communicability of STD from man to woman and β: that of STD from woman to man. Here, p(k) is defined as the probability distribution of men in a society who have k sex partners and q (k), that of women.
Regression analysis was attempted for data on daily mortality from circulatory diseases and daily weather conditions in Tokyo from 1980 to 1981. The relation between the mortality and weather conditions was also examined. For representing the mortality by weather variables, several linear regression models and several autoregression (AR) models including weather variables as common factors were predicted, and applied to the data in 1980. The models identified by the data in 1980 were used for predicting the daily mortality in 1981. Validities of the models were evaluated by this procedure. The latter models generally fit the data better than the former models. Actually, an AR model which included only the daily minimum air temperature as a common factor was selected as the best model among our models. We emphasized that the daily minimum air temperature is closely related to the mortality, although this fact has already been reported in various studies. Further, we found that the variations of air temperature or vapor pressure affected the mortality. This fact was shown by residual analysis. It was also noted that the seasonal features of the residuals decreased in winter and increased in spring because of the difference of weather stability.
This study examines divergent perceptions of shoreline quality among landowners of the Adirondack region of New York State. The social influences behind these perceptions fit into three groupings: personal background, community characteristics and public activities affecting shoreline quality maintenance. The relative effects of these influences are assessed through correlation, regression and multidimensional scaling method on a random sample of 713 property owners. Government policies, especially land use regulations, are shown to be a major influence on perceived shoreline quality once community and personal background are controlled.
Here proposed is a numerical method for distinguishing some divergent series from convergent ones. It is applicable to numerical integration of some functions over finite domains by iterative application of trapezoidal formula, as well as to some basic series in higher education, such as the harmonic series, ordinary Dirichlet series, and so on.
It is difficult to compare several categories at the same time. The method of paired comparisons has some advantages when a fine judgement is needed. As the method of paired comparisons in addition to the Bradley's paired comparisons, various procedures are hitherto used. However, the traditional method of paired comparisons may be regarded as a special rank order technique. This paper proposes the procedure based on the non-rank paired comparisons (RPO-research) different from the traditional procedures. The method of analysis based on RPO-research is called the RPO-analysis for short. It is a practical method of analysis to compute the solutions of the system of linear equations concerning the Relative Proportion of Occupation. In an application, it is explained that we can know the invisible consciousness of students by the result of the RPO-analysis on 1600 students and their parents.
The purpose of the present study was, first, to examine the preference patterns of different types of decision strategies — additive (ADD), additive difference (ADD-DIF), conjunctive (CONJ), lexicographic (LEX)— in the decision task with different numbers of alternatives and attributes, and secondly, to investigate the relationships between the preferences of decision strategies and the indexes of the decision making process. The subjects were sixty-seven female and male university students. The main results were as follows: (1) Although there were interrelationships among the preferences of decision strategies, the preference pattern of each strategy was consistent among the different decision task situations. (2) The preference of each decision strategies was more determined by perceived optimizing potential of the strategy than perceived simplifying potential in the decision making process.
Log-linear models have played a central role in analyzing the contingency tables. These models fail to exploit all the available information about the category orderings. Goodman (1979, 1981a, b) has recently proposed a general class of models for the analysis of association in cross-classifications having ordered categories. Clogg (1982), Agresti (1983) and Agresti & Kezouh (1983) extended his method to multi-way cross-classifications. This article presents the maximum likelihood methods to obtain parameter estimates for the analysis of association in multi-way case.
Many multivariate techniques such as principal component analysis, dual scaling and discriminant analysis can be formulated as eigenvalue problems, including generalized eigenvalue problems. This paper introduces a sensitivity analysis method developing the idea of W.S. De Sarbo, et al., who proposed the use of an ε-neighborhood around the maximum value of the criterion function. They and W.J. Krzanowski derived an approximate ε-neighborhood. This paper presents the exact ε-neighborhood. As application examples of the exact ε-neighborhood, dual scaling for conditional ordinal data and principal component analysis are discussed.