Data analysis means here data processing to obtain the spatial representation of structures hidden in data without external criteria. The former part of this paper first presents a general view on current methods in data analysis, and then in contrast to quantification theories discusses how to utilize multidimensional scaling(MDS)methods as models accounting for data or tools of data analysis. The latter part raises and exmaines two problems in practical applications of MDS. The one is the indeterminancy of solutions inherent to models. The other is the possible consequences of using simplified models in comparison with the form of observed data.
This study was purposed to examine the new possibility of an application of Multidimensional Scaling Method to the study of mass media function. We considered television, radio, newspapers, magazines, books, and movies as six important media of mass communication, and analysed how they function and gratify various“needs”in our daily life. Facet Theory by Guttman was used to construct a certain psychological space of the individual. MDSCAL produced some interesting results, and we discussed about the relationship and inter-changability of these six media.
In this paper, I try to apply the PATTERN method(Planning Assistance Through Technical Evaluation of Relevance Number)for the decision process in the public system and attempt to construct a theoretical framework to explain the PATTERN method procedures.In this way, the PATTERN method procedures are to understand the public problem and solve it by three steps as follows.(a)To write the scenario from the public's point of view.(b)To build up the relevance tree from the content of the scenario.(c)To make the quantitative evaluation on the relevance tree.From these steps is developed a theoretical framework to construct the relevance tree from the scenario writing. At the same time, the multistage decision theory is applied to this relevance tree through the quantitative evaluation.This method is obtained by the consensus of the public opinions. Meanwhile, a decision making can be made in the public problem.Finally, this method is also applied to perform a case study of“the Public Problem of the Industrial Waste Treatment Planning”. I made a summary of it and examined its result.
The models proposed here can be classified into three groups; a)models of communciation networks, b)models of balancing processes, and c)a model of social power. The first group are descriptive and treats with changes in frequencies of communciation in a group. The second group consists of two subgroups: one is direct generalization of Flament(1963)'s model with respect to the probabilities of transition and the other allows for the neutralization of opinions. The third model is generalization of Harary(1959)'s model and its appcliation to Jacobs and Campbell(1961)'s experiment is presented as an example of its uses.
The anticipation of the time of parturition was tried by the multiple regression based on items which indicated the advance in pregnancy toward parturition. Four multiple linear regression function of log T were prepared for anticipation in primiparous women before and after the rupture and in multiparous women before and after the rupture, respectively, where T express the interval between the time when each item was measured and the time of parturition.Parturition Process was very different for each individual. But by the similar method as the Covariance Analysis, it was ascertained that the difference was suffciently expressed by constant which vary with each individual and linear coeffcient were common to each individual.
The face graph proposed by Chernoff has much contributed to grasp the global structures of the multi-dimensional data in many fields. However, when the data of one's body such as stature, body-weight, chest-circumference and sitting-height are concerned, a body graph proposed in the present paper can represent the global structures of the data more efficiently than the face graph. In this paper we mention how to construct and to use a body graph through some practical examples.
During the past decade, the number of published studies on scaling and related topics has increased at an explosive rate. This article provides bibliographic information of this general area, with reviews of research trends within and across the following categories of research: unfolding techniques; paired comparisons, ranking, and seriation; optimal scaling; conjoint measurement scaling; multidimensional scaling; homogeneity analysis(clustering, pattern and profile analysis, etc.); psychophysical scaling; theories of measurement; other scaling techniques and related topics, and; computer programs. The present article deals with the first four categories. The bibliographies, which cover mainly the past ten years, are by no means exhaustive, or rather, may not include half the studies published during the period. In particular, most studies published in Japan are not included.